v>EPA
820-R-10-018
Fluoride:
Dose-Response Analysis
For Non-cancer Effects
Dental Fluorosis:
Evaluations of Key Studies
Health and Ecological Criteria Division
Office of Water
January, 2008
U.S. Environmental Protection Agency
Washington, D.C.
-------
TABLE OF CONTENTS
ACKNOWLEDGMENTS 3
INTRODUCTION 4
STUDY SUMMARIES 5
January, 2008
-------
ACKNOWLEDGMENTS
This document was prepared by Oak Ridge National Laboratory, Oak Ridge, Tennessee, under work assignment
2006-014, underthe U.S. EPA IAG Number DW-89-9220971. The Principal EPA Scientist is Joyce M. Donohue,
Ph.D., Health and Ecological Criteria Division, Office of Science and Technology, Office of Water, U.S.
Environmental Protection Agency, Washington, DC.
The summaries included in this report were prepared by C. Wood, S. Milanez, D. Glass, S. Garcia, S. Goldhaber,
and V. Dobozy. Summary reviewers included J.M. Donohue and T. Duke of the Health and Ecological Criteria
Division, Office of Science and Technology, Office of Water, U.S. EPA; and D. Glass, D. Opresko and A. Watson
ofORNL.
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.
January, 2008
-------
INTRODUCTION
Prior to initiating the dose-response analysis for severe dental fluorosis, the Office of Water (OW)
critically evaluated the studies that had been cited and utilized by the National Research Council (NRC,
2006) in their report Fluoride in Drinking Water: A Scientific Review of EPA 's Standards. Additional
studies identified in the OW initial literature search (2006) were also evaluated. Critical information
fields examined and summarized include endpoint studied, type of study and population studied,
exposure period and assessment, characterization of study groups, analytical methods and study design,
parameters monitored, statistical methods employed, results (including critical tables and figures)
authors' conclusions, critical references and definitions, profiler's appraisal, and critical review of the
profiler's assessment.
This document is a compilation of the study evaluations arranged alphabetically by the name of the lead
author. Dental fluorosis studies identified and added to the dose-response analysis for the non-cancer
effects document after its external peer review were not evaluated in this fashion.
January, 2008
-------
STUDY SUMMARIES
Dental Fluorosis
January, 2008
-------
Acharya, S. 2005. Dental caries, its surface susceptibility and dental fluorosis in South India.
International Dental Journal, 55(6): 359-64.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
Dental fluorosis and dental caries
Cross-sectional survey
India/Karnataka State (Deccan Peninsula). 544 schoolchildren (301 males and 243 females)
aged 12-15 years old from five different villages (Nallur, Naganur, Doddabathi, Kundawada
and Holesirigere) within Karnataka State (Davangere District) were studied. All children were
native to the region.
none
Schoolchildren that were continuous residents of the studied villages since birth and ranged in
age from 12-15 years.
The number of children examined in each village and the fluoride concentration identified in
the water supply for the village are provided in the table below. All children were in
approximately the same socioeconomic status and sorghum was the main staple food item.
Water was obtained in the villages from bore wells and was either pumped to a centralized
storage tank where residents obtained water through a tap on the tank or through the
Accelerated Rural Water Supply which distributed the bore well water to the consumers. The
wells were approximately 15-18 years old.
Name of Village No. of children examined Fluoride cone, of water
Nallur 163 0.43 ppm
Naganur 49 0.72 ppm
Kundawada 96 1.10 ppm
Doddabathi 81 1.22 ppm
Holesirigere 155 3. 41 ppm
Only drinking water was analyzed for fluoride content in this study. The fluoride content in
food and beverages (i.e. tea) was not measured. Information regarding food habits was asked
of the participants but data characterizing consumption or fluoride concentrations were not
included for the assessment.
Information about the age of the wells (15-18 years old) and the fact that the water supply had
been from a constant source was obtained from the local and village councils and the Public
Health Engineering Dept. of Davangere District. The analytical method used to measure
fluoride in the wells was the ion selective electrode method developed by the Orion Research
Incorporated Laboratories Products Group, USA. The model used was (94.09, 96.09)
electrode 720 A from Orion Instruments. Information about other parameters measured in the
water was not included in the paper.
544 schoolchildren (301 males and 243 females) aged 12-15 years old fromfive different
villages (Nallur, Naganur, Doddabathi, Kundawada and Holesirigere) within Karnataka State
(Davangere District) were studied. The age distribution was 170 (31.2%) in the 12-13 year old
group, 201 (36.9%) in the 13-14 year old group and 173 (31.8%) in the 14-15 year old group.
Dental examinations were performed one time on the children in their schools. Children were
examined under natural light while sitting on a chair or stool. Children were assessed for the
presence and degree of fluorosis and the evidence of caries.
The author of this paper (S. Acharya) was the dental examiner for the study and was trained in
the WHO criteria for assessing fluorosis and caries prior to the start of the study. Dean's
January, 2008
-------
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
Dental caries
criteria (1942) were used to assess fluorosis, and the Community Fluorosis Index (CFI) (Dean
1942) was calculated to assess the public health significance of fluorosis from each village.
Dental caries were assessed using the DMFS index (Klein et al. 1938). The type of carious
lesion present was also recorded: pit and fissure lesions or smooth surface lesions. Occlusal,
lingual and buccal pit and fissure lesions were classified under pit and fissure lesions.
Proximal lesions and lesions on buccal, lingual and occlusal and incisal surfaces other than pit
and fissures were classified under smooth surface lesions.
All data analysis was done onMinitab Statistical Software (Version 13). The Karl Pearson
coefficient for correlation and simple regression analysis was used to measure the correlation
between fluoride concentration in the drinking water and dental caries. The F test was used for
estimation of statistical significance and statistical significance was considered when p< 0.05.
A sub-sample of 10% of the schoolchildren was re-examined for fluorosis and dental caries
with one-day intervals between examinations to assess intraexaminer variability; Cohen's
Kappa coefficient was found to be 0.88, indicating a high level of agreement.
Table 2 is copied directly from Acharya (2005) showing the prevalence and severity of
fluorosis associated with exposure to water with varying fluoride concentrations. The
prevalence of fluorosis increased from 16% at 0.43 ppmF to 100% at 3.41 ppmF and the
degree of fluorosis severity increased as fluoride levels increased. The Community Fluorosis
Index (CFI) increased from 0.10 at 0.43 ppmF to 2. 10 at 3.41 ppmF, making the 3.41 ppm
community one of marked health significance.
T®bS0 2 Prevaltne© ami stvwity 01 ffuorssis in relation to diSf ring water fluoride levels
F tev^I Total msmtw Ruorosis Sdwrity of ftuofosis CFI
(ppm) ot cases prevalence Normal Questionable Very mlW Mik) Moderate Severe value
N % n % n % n % N % n% n %
0.43 163 26 n 137 84 21 12.8 3 1.8 2 1.2 00 00 0.10
0-72 49 K 51 24 49 8 18.3 10 2.04 8 10.2 1 0.6 00 0,56
1-10 88 54 58,2 42 43.7 26 27 15 18.8 S 9,3 4 4,1 00 0,80
1.22 81 44 54,3 37 45,6 24 29.8 9 11.1 7 8,6 4 4.8 00 O.SS
3'*t 1S5 155 100 00 S 3,22 81 39.3 26 18.1 43 Z7.T it 13,5 2.10
ToM 544 304 SS.8 840 44,1 85 15.8 9S 18 30 .5 52 8.5 21 3.86
PROFILER' S NOTE: The prevalence and severity of fluorosis is given for each village was
assessed; however, the greatest variance in data were when fluoride levels went from 1.22
ppm to 3.14 ppm. Acharya (2005) discusses the high level of fluoride present within this
region of India due to the use of phosphate fertilizers causing fluoride levels to be increased in
foods such as sorghum and rice (Anasuya et al. 1997) and the heavy consumption of tea in
India; however, actual consumption of water or these other fluoride sources were not
addressed in the study.
Acharya (2005) states that caries incidence and severity were highest in the children living in
the area with the lowest fluoride (0.43 ppm) concentration, and that there was a statistically
significant negative (r = -0. 16) correlation between water fluoride levels and the mean DMFS,
showing a declining trend with increasing level of fluoride (see Table 1 and Figure 1 copied
directly from Acharya, 2005). The pit and fissure lesions also showed a decrease with
increasing water fluoride levels, however, this trend was not observed with smooth surface
lesions. Dental caries also was stated to be more prevalent in the older children and in the
females.
January, 2008
-------
Caries surface patterns
STUDY AUTHORS'
CONCLUSIONS:
Table 1 Csrres prevalence snd patterns m relation t® differing Hyorkfe
laveis
f Level Number of eases €,««»;, prwileno Mr mf'MFS SDi
Cppm! n <*„
0.43 163 86 f? C
0,72 49 26 531
1.W % 38 39 fi
1,22 81 30 TO
3.41 1SS 5* d4 8
F.5.S7, peO.001 (US)
SO = Slandard Deviation
2.0
W 1.6 .
1 1'2 *~~~*~~~^~~~~~~^^_
| 0.8 «" ~
I 0.4
0.0
0.6 0.5 1.0 IS 2.0 2.5
FI.Level(ppm)
Flgum 1. CamMton and regression between water fluoride levels and mean DMFS valu
iqyation: 1 .42 - 0.22 (F level). An inverse correlation was seen betwmn water lluoride 1
slfniteanl
PROFILER'S NOTE: While the prevalence of caries did decre
increased, the trend was not as strong for the DMFS score. The
1.10 ppm fluoride level group was actually less than that at 1.2.
The biggest difference in caries prevalence and mean DMFS sc
highest fluoride levels were compared. From the data provided
that the trend of increasing dental caries with age and in female
1 S8l 98 1
1 r0t 4Si
T8S, 34,
0 QJI (,?'.
t 7«i 39t
Prt dietwl DMFS ^
Actail DMFS _
~ -_»
3.0 3.5 4.0
3s. CoTOtafcn Coefficient; - 0.16; Regression
svels and msan DMFS which was statistically
;ase as the fluoride level
mean DMFS score for the
1 ppm (see Table 1 above).
ores was when the lowest and
the profiler could not agree
s occurred.
The number of pit and fissure lesions as compared to smooth surface lesions are shown in
Table 3 copied directly from Acharya (2005). The number of pit and fissure lesions decreased
as the fluoride levels increased.
Table 3 Cartes surface patterns in relation to differing water liuoride levels
F Level Number of cases Mean number of pit yean number of
fppm) Mid tissue iesmns smooth surfae© lesions
0.43 163 1.34 0.22
0.72 49 t 16 0.04
1.10 98 0.72 0.16
1 .22 81 0 88 0.0?
3.41 155 067 0.07
Mean DMFS showed a decreasing trend with increasing fluoride levels in the
live villages. Pit and fissure lesions showed a definite decrease with increas-
ing water fluoride levels.
Of the 544 children examined, 234 had caries; however, only 10 filled tooth surfaces were
found, indicating insufficient oral health care in the region.
There was a highly significant negative correlation between water fluoride levels and dental
caries. Dental fluorosis also increased with increasing fluoride levels. Caries surface patterns
were defined as either pit and fissure or smooth surface lesions, with pit and fissure lesions
more common than smooth surface lesions. Pit and fissure lesions showed a decreasing trend
with increasing fluoride levels but this trend was not observed with smooth surface lesions.
Overall, water fluoride was an important factor associated with low caries prevalence.
Acharya (2005) also observed that low caries and small numbers of smooth surface lesions
occurred in the teeth of children from areas with low fluoride concentrations in drinking
water. Acharya hypothesizes that this may be due to the "very active local economy" of food
produce and food items, in which interexchange of food produced (either grown or cooked) in
January, 2008
-------
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S DFG/12-06
REMARKS and
12/14/2006
PROFILER'S ESTEM.
NOEL/NOAEL
PROFILER'S ESTEM.
LOEL/ LOAEL
POTENTIAL
SUITABILITY FOR DOSE-
RESPONSE MODELING:
CRITICAL EFFECT(S):
communities with high- and medium-fluoride wells is frequently bought, exchanged, sold and
consumed in communities supplied by low-fluoride water wells. Acharya (2005) considers
this exchange to result in a "'halo' effect." Consumption of tea (considered a considerable
source of F), salt and other condiments may also be factors.
Anasuya, A., S. Bapurao and P.K. Paranjape. 1997. Fluoride and silicon intake in normal and
endemic fluorotic areas. J. Trace Elements Med. Biol., 10:149-155.
Klein, H. C.E. Palmer and J.V. Knutson. 1938. Studies on dental caries I. Dental status and
dental needs of elementary school children. Public Health Rep., 53: 751.
Data showed a dose-response to fluoride concentrations in the water as the severity of
fluorosis; the CFI increased when fluoride concentrations went from 1.22 ppmto 3.14 ppm.
For fluoride concentrations between 0.72 and 1.22 ppm, there was not much of a dose-
response and the health significance based on the CFI was borderline. The study did not
present the data based on age levels or gender so the profiler could not confirm all of
Acharya's (2005) conclusions. While Acharya (2005) discussed the other possible sources of
fluoride that have been documented in the area, there were no data provided to characterize
them. The study did use accepted standards of assessment such as Dean' s index for fluorosis.
Statistical analysis appears to be adequate.
The study also indicated a decreased amount of caries as the fluoride levels increased in the
water.
This study was not designed to be suitable for development of a NOAEL for fluorosis.
This study was not designed to be suitable for development of a LOAEL for fluorosis.
Not suitable (_), Poor (X), Medium ( J, Strong (_ )
Data showed increased severity of fluorosis and an increased CFI when the fluoride levels
went from 1 .22 to 3. 14 ppm but there was not much difference in those exposed to 0.72 to
1.22 ppm making a clear dose-response not evident. In addition, the authors note likely
confounding due to dietary fluoride intake by villagers who are also supplied with low-F
drinking water. Therefore, the study does not have a potential to be used for dose-response
modelling.
Dental fluorosis and dental caries
January, 2008
-------
Akpata, E.S., Z. Fakiha and N. Khan. 1997. Dental fluorosis in 12-15-year-old rural children
exposed to fluorides from well drinking water in the Hail region of Saudi Arabia. Community
Dent Oral Epidemiol 25:324-7.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
Dental caries and fluorosis (permanent teeth).
Prevalence study.
Saudi Arabia/Hail region: Children aged 12-15 years; selected for study participation by a
two-stage stratified cluster sampling technique. Hail villages were stratified according to the
fluoride concentrations of the wells used (see Table 1). Each stratum of fluoride
concentration was allocated a sample size proportional to its population. Forty-two primary
and intermediate schools, chosen by simple random sampling technique from a total of about
155 were visited and a sample of classrooms with children aged 12-15 years was selected by
the same technique. The 2355 children (approximately equal numbers of boys and girls)
examined were life-long residents of the villages and had no obvious nutritional deficiencies.
Those who obtained their drinking water from more than one well were excluded from the
study. Informed consent for study participation was obtained from the community heads and
school authorities.
None.
12-15 yr; from time of birth to!993.
0.50-0.79; 0.80-1.09; 1.10-1.39; 1.40-1.69; 1.70-1.99; 2.00-2.29 or >2.30ppm fluoride in
well water. Of the 1083 wells in the Hail region of Saudi Arabia, 87 popularly used for
drinking water were selected for analyses. The wells were at least 20 years old and about 300
m deep. The wells contained ground water, possibly located at the confined aquiferous earth
stratum and consequently, significant seasonal variation in the fluoride concentration of well
water was unlikely. The study authors note that the amount of time the children spent in air-
conditioned rooms might have influenced water drinking habits, and thereby affected
fluoride intake; however, this factor was not quantified.
The study author note that the data suggest other sources of fluoride exposure such as foods,
beverages and infant formula, although fluoride intake from these sources was not
quantified.
Fluoride levels in the 87 wells selected for analysis were measured by the ion-specific
electrode method. (Taves, 1968)
The objective of the study was to investigate the relationship between fluoride levels in well
drinking water, severity of dental fluorosis and dental caries in the Hail region of Saudi
Arabia. A random sample of 2355 children aged 12-15 years was examined for dental caries
following the WHO criteria (WHO, 1977). Their teeth were then examined for dental
fluorosis using the Thylstrup and Fejerskov Index (TFI; see NRC, 2006, pages 88-89). A
reproducibility test was conducted on 20 subjects to test for intra-examiner agreement.
Dental caries was evaluated following the WHO criteria. Dental fluorosis was evaluated
using the modified TFI. The teeth were dried with gauze, illuminated with a pocket torch,
and the facial tooth surfaces examined for fluorosis. To calibrate the three examiners against
the principal investigator, 20 subjects were examined for dental caries and dental fluorosis of
tooth #11. The calibration was repeated until intra-examiner level of agreement gave a
Cohen's Kappa statistic of at least 0.75. At each session during the field work, about 10% of
subjects were examined a second time and the reproducibility measured; this resulted in
intra-examiner level of agreement varying between 0.76-0.80 and 0.78-0.84 in the diagnoses
of dental fluorosis and dental caries, respectively.
10
January, 2008
-------
STATISTICAL METHODS:
Chi-square tests, simple linear regression analysis and analysis of covariance were used to
assess the statistical significance of the association between fluoride levels in well drinking
water, severity of dental fluorosis and caries occurrence.
RESULTS:
The study population is presented in Table 1 taken directly from Akpata, 1997.
ti'n finm i
O.JO 0
(I W 1
f HI I
2 30
JOftt
! W
'fit
i.'B.
Study results in Tables 2-3 and Figures 1-3 are shown directly from Akpata, 1997.
t*
IS
n i »
fUDCL
Itrtl*
2 iff-
M
fi ?J
0 !li
ft I'-
ll .'
f
n i:
0 13
0 10
n\fFf!
J.73 r 7
_t (hi - T-1
p.
21*5 K«I chi.J
r*pOK<
i<
*>
ts
413
2 MS
4*
v'J
Ml
11
January, 2008
-------
- TFI «»
- T»I a*
i j
•r.
,,. ,,A-
I 4»
120
100
_
!!?-§ '.35-6 • 13-4 I-;1 1-2 t :0 :
11
j..» ^
Tooth Type
drtinltrfmrt rtt "II 1 worm >it
li- f; 5
n> wdl
tooth type* of I2-IS-
12
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/Date
REMARKS VAD/03-09-
07
PROFILER'S ESTIM.
NOAEL
PROFILER'S ESTIM.
LOAEL
Mi ' •• - • — __
:fi?~9 Cls-6 I
" I
' "" ,,_ f
3-4 "l>2 f.lC
«,m « . ,
k™™^**™™*^^^
, .<, ™, > -fy -t™ a-w-w. -M taa™ | r- m ^ir in.~^gp -;— — w — .::i. Mi'n ^>; ~ ~T- - j
f3 *s n \y V^ ^ ' lAVyX^'v
5
i
20 ;
•' . ' ,;'
i _ * 1-
i_ "^f
1
. ... _ " ' * "?
,..,: J '•'.,. i
" 1' U » ' " > „ « -.,1.
1 - '~*:M t in Hwwiibiaiat loath types of J2 1*"
Wi»r-p!«t M writ wastr.
The study authors concluded that the strong association between fluoride levels in well
drinking water and the severity of dental fluorosis indicates that fluoride from well water is a
significant factor in the etiology of dental fluorosis in the Hail region. Although there was a
statistically significant relationship between DMFT (decayed, missing or filled teeth) and
fluoride levels, the R2 value was very low, indicating that fluoride in well water made little
contribution to the variability of caries experience in Hail children.
World Health Organization. Oral health surveys: basic methods. Geneva: WHO, 1977.
Taves, D.R. 1968. Determination of sub micro molar concentrations of fluoride in bio logical
samples. Talentia 15:1015-23.
The study results are not representative of the U.S. population since the study was conducted
in Saudi Arabia. The study didn't account for other sources of fluoride exposure in addition
to the drinking water. The study report indicated that attempts have been made within the
past decade by some of the Hail rural population to defluoridate their well drinking water by
sedimentation and distillation. These practices may have reduced the actual fluoride
concentrations in the drinking water of some children in the study.
Although there may have been confounding factors, such as additional fluoride intake
through food or beverage consumption, the data do show a trend towards increasing severe
fluorosis with increasing fluoride concentration in well water (about 6% at 0.5-0.79 ppm to
greater than 20% at 1 .40 ppm and higher. Although the incidences of severe fluorosis are
higher in this study, comparisons can be made to the Galagan and Lamason (1953) study
from the southwest US where similar climatic conditions occur.
The study design did not identify a no-flurosis intake dose.
Fluorosis, including severe fluorosis, was reported at all levels of fluoride concentration in
the drinking water. Therefore, the LOAEL was 0.50-0.79 ppm.
13
January, 2008
-------
SUITABILITY FOR DOSE
RESPONSE MODELING
CRITICAL EFFECTS:
Not suitable,O; Poor Q; Medium (X); Strong (_)
Dental fluorosis and caries (permanent teeth)
14
January, 2008
-------
Awadia, A.K., Birkeland, J.M., Haugejorden, O., and Bjorvatn, K. 2000. An Attempt to Explain
Why Tanzanian Children Drinking Water Containing 0.2 or 3.6 mg Fluoride Per Liter Exhibit
a Similar Level of Dental Fluorosis. Clin Oral Invest 4: 238-244.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
Dental fluorosis
Cohort
Africa, Tanzania: 80 school children of African ethnicity, ages 8-16 years old, from the
urban community of Arusha, Tanzania where the fluoride level in the drinking water is
3.6mg/l.
96 school children of African ethnicity, ages 8-16 years old, from the rural community of
Kibosho, Tanzania where the fluoride level in the drinking water is 0.2 mg/1.
A structured interview was conducted to collect information relating to the child's first 6
years of life. The enamel formation of upper central incisors is normally finished at
approximately 3.5 years after birth. Thus, the use of tooth 2 1 (upper left central incisor)
in the analyses restricted relevant fluoride exposure to a limited period of life.
Subjects were from two northern Tanzanian communities. Kibosho (altitude 1300 m) is a
rural community of approximately 20,000 inhabitants (mostly of the Wachagga tribe with
African ethnicity) and has a historical water fluoride concentration of 0.05-0.56 mg/1.
Arusha (altitude 1400 m) is an urban community of 135,000 inhabitants (multi-ethnic and
multi-tribal) and has a historical water fluoride concentration of 3.5-3 .6 mg/1. Data is
based on findings over 20 years.
Children were selected from four schools in Arusha (every third child was randomly
selected from the schools' attendance records, n=80) and from one school in Kibosho (all
children in grades 3-5, n= 96). The average age was significantly lower in Arusha
(10.4±1.8years) than in Kibosho (12.4±1.7years). There was no gender difference
between or within areas.
Magadi, a fluoride-containing salt added to weaning food and to 'adult' food, may be an
important source of fluoride in Kibosho. On a dry weight basis, a magadi sample has
been shown to contain 1.5 mg F/g. An estimated intake of 0.44 g magadi per adult per
day would give a fluoride exposure of 0.7 mg/day. However, there is marked variability
in fluoride content of magadi (160 to 1750 mg/1). The amount and the fluoride content of
consumed magadi could not be assessed in the current study.
Dental fluorosis prevalence and severity were measured; comparisons were made on TFI
scores from tooth 2 1 . Information regarding variables assumed to be related to dental
fluorosis during the first 6 years of life was collected from questionnaires consisting of
two parts; the subjects answered one part and the accompanying parent(s) (mainly
mothers), the other part.
Data on how fluoride concentrations in the water supply were measured were not
included in the study report. Fluoride content in magadi was not determined in the study.
The study included two cohorts of school children from two communities in northern
Tanzania. One study population was from Arusha (n=80, age 10.4±1.8 years), a multi-
ethnic and multi-tribal urban community with a water fluoride concentration of 3.6 mg/1.
The other study population was from Kibosho (n= 96, age 12.4±1 .7 years), a rural
community consisting primarily of Wachagga tribe members with African ethnicity and
with a water fluoride concentration of 0.2 mg/1. Children were selected from four schools
in Arusha (every third child was randomly selected from the schools' attendance records,
n=80) and from one school in Kibosho (all children in grades 3-5).
15
January, 2008
-------
Information regarding variables assumed to be related to dental fluorosis during the first 6
years of life was collected from questionnaires consisting of two parts; the subjects
answered one part and the accompanying parent(s) (mainly mothers), the other part.
Variables included: fluid intake; diet, including weaning food (i.e., lishe—porridge
containing ground maize, beans, peanuts, and fishmeal; and kiborou—beans and bananas
cooked with magadi, a fluoride-containing salt); toothpaste use; parent occupation; and
general health. The prevalence and severity of dental fluorosis was measured as follows:
Examinations: Subjects were examined for dental fluorosis under indirect light while
seated on a chair. Prior to the clinical examination, the teeth were wiped clean with
cotton gauze and isolated with cotton rolls. One examiner graded dental fluorosis on the
facial surfaces of all permanent teeth using the Thylstrup and Fejerskov Index (TFI).
Substantial intra-examiner agreement of the recordings has been reported earlier (Cohen's
kappa 0.74). To standardize conditions and allow comparison with previous findings, the
TFI score for tooth 21 (upper left central incisor) was used to characterize the severity of
fluorosis in a subject. No radiographs were taken during the surveys.
PARAMETERS
MONITORED:
Dental fluorosis was measured using the Thylstrup and Fejerskov Index (TFI). No
radiographs were taken during the surveys. The TFI score for tooth 21 (upper left central
incisor) was used to characterize the severity of fluorosis in a subject.
STATISTICAL METHODS:
Data were analyzed using the statistical package for social sciences (SPAA for PC version
9.0). Distribution of subjects according to background variables was assessed by chi-
square tests (with Yates's continuity correction). The average total fluid intake in the two
areas was compared using the Student's /-test. The Mann-Whitney [/test was used to
compare the distribution of fluorosis between areas and background variables.
Spearman's rank correlation coefficient (rs) was employed to assess the strength of the
bivariate association between the dependent variable (TFI score on tooth 21) and the
independent variables. Stepwise multiple linear regression analyses (SMLRA) was
applied to control for confounding and to estimate explained variance in the dependent
variables. The SMLRA was run for the whole group (inter-area) and within areas. The
intra-area analyses were carried out in an attempt to further explain the difference in the
level of fluorosis between the areas. Variables logically important for the development of
dental fluorosis were selected for the regression model rather than variables found to be
significant in bivariate analyses. The level of significance was 5% (p<0.05).
RESULTS:
Dental fluorosis
Figure 1 was copied directly from Awadia et al. (2000) and summarizes the frequency
distribution of children according to the severity of dental fluorosis, recorded on tooth 21.
The prevalence (TFI >1) was not significantly different between the communities. The
median TFI score on tooth 21 was four in both areas. The severity of dental fluorosis was
significantly lower in Kibosho (0.2 mg F/l) than in Arusha (3.6 g mg F/l) (p=0.008).
There were no significant differences between genders.
16
January, 2008
-------
100%
40'«
20%
0"n
Arusfta
iTFla D, BTf :*4. QTFi: 2, Z
Fig. 1 Frs^'jew d,s:r.!r.i&Qi: V af rbl-drea ?c:cr'iii| is the
3j'm™-' ef i-Esl flac:a:j» recoriel iy foe T7 be;1! 02 taot". 11 c
Il'.tciiij il'_ EE £jtnd«l, ,'>=£"? }j£ Arias ,5 ? 3:1 f.'jor.de I
Figure 2 also was copied directly from Awadia et al. (2000) and summarizes the
distribution of children and median TFI scores on tooth 2 1 according to area and dietary
practices. When comparing dental fluorosis among magadi users (comparison^ in
Figure 2), the scores in Kibosho (TFI=4, n=94) were significantly lower than in Arusha
(TFI=5, n=35, p=0.002). The TFI scores among users of both magadi and kiborou in
Kibosho (TFI=4, n=36, comparison B in Figure 2) did not differ significantly from the
severity in magadi non-users in Arusha (n=44, p=0.93).
..
K£ IEJ±.IJI .rl-.:3re>ontoc:t
"- '. accor±ii| :a JK: .ml Jaucy
^ra"i;a; ia'!l:'3Cil'j ;C 2 eg
Biicr.ca j. tf=?l, nat Aniin:
,i 5 a;| P 1
-------
STUDY AUTHORS'
CONCLUSIONS:
Srac.rJS'fkiSJ'" Erdigmaiv™^ C«fKr :::-^CL%> ^h:,%- 7K
"ilcJir _-_ _' _l 1 J
,, „ TniaH.LK-.s' -11.C.1V t.'iTf >,*"*) I-l
, -\5? .",? '. *-e \"*i» '..Say 44' .*. ,j i i-l
1 liJ;:Zi- j^ni L3T 3'. *i M la " ~rii>ca)<; •''':• i- ?"
lirar.- i. 'i- ~x'*,i' •* -' \~a " -?' " * •»'
9-0 'j=c'«' fe 1 ;jLi-..eT " -L-*, ^ 4^ ""•",--, l'"~
r.^' jiratfcrLj A
Table 4 was copied directly from Awadia et al. (2000) and summarizes Spearman's rank
correlation coefficient between TFI score on tooth 21, age, dietary practices, habits and
liquid intake (n=171-176). Area and fluid intake were significantly associated with the
TFI score inbivariate correlation analyses. Area correlated significantly with several
independent variables. Use of magadi was significantly associated with mother's
occupation (peasantry) and hence with area of residence in the bivariate analyses.
Table 4 '..jearciir. rr.rj. csrre'mor. ,"«5:ieit CKWME IFE SCOT? or. csofr 21. is* omcy prjcac*1.. ku.K md l:y
';=171-iVf. ff^jr E-oaeaatep«ntkS4.1i'i>£ac-:l'.»r ; acirjpaLct rjfjtw cJcc-.qKneu
TFIscc^ ATS: A|e Hxp ?*."ad XO FO T-ps;:e K;
Are; - 20"
Ass 12 0 "'2"
F'fcid icsie * -04$" -1 If -Ol5*
MC - Oo15'* o-}'** iVJ -?22"
F^ - 0 >* } .S M Uit -.1 .5 -*
.30!^pi^S^:>5 -HJ _5 * ™0 - 1 ~^! . "^ _. — ^ '„' -. _^S
E»?Ii-a,5 --Jli'f -305 -OOS -114 -521*' -•) 15 C 20'* -0.
Are? 3=Air.r!;:,, 1= K osio. t«a,nre Ct=criec belt; l=sr^?a ,fr«iln MEJ;
•f oo; -*f ox '? ooc: i.'-"-*-.-(
Table 5 was copied directly from Awadia et al. (2000) and summarizes the stepwise
multiple linear regression analyses with TFI score on tooth 2 1 as the dependent variable
(n=176). When the variables age, area, total fluid intake, magadi use, weaning food use,
and mother's occupation were used in the SMLRA, the area accounted for 3.2% of the
variance in the TFI score. The use of magadi increased the variance that could be
accounted for by 2.9% when the total explained variance was 5.0%. InKibosho, age
explained 5.7% of the variance in TFI score in SMLRA (p=0.02). In Arusha, magadi use
significantly explained 3.7% of the variance in TFI score.
F,="1- ^'T.CO^S. aoi uT-ji e^.p'.jL iLi.i 1*. As'LLi fajJ i?=i ic iE-,P=v!.> a. j • .r:i > f =
-------
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date
REMARKS SJG/2/14/07
PROFILER'S ESTEM.
NOEL/NOAEL
PROFILER'S ESTEM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
The fluoride concentration of the drinking water (18-fold), degree of urbanization and
tribe were the obvious differences between areas. The participants differed significantly
regarding practices related to food habits such as boiling of drinking water, use of magadi,
and type of weaning food. Except for area, use of magadi and weaning food, none of the
above mentioned factors explained a significant proportion of the variance of the TFI
scores in the multivariate analyses and total explained variance was only 5%.
Given the different fluid intake levels (0.7 in Kibosho vs. 1 .0 1 in Arusha) and the fluoride
concentrations (0.2 vs. 3.6 mg F/l) in the drinking water, the fluoride intake in Arusha
corresponds to approximately 3.5 mg F/day. Thus, alternative sources of fluoride are
required to explain the level of fluorosis observed in Kibosho. Magadi users had
significantly higher TFI scores in Arusha. Statistically, the effect of magadi could not be
established in Kibosho since only two subjects were non-users. Magadi may be an
important source of fluoride in Kibosho. Whether the use of magadi and/or the traditional
weaning food (kiborou) is primarily related to tribe, area, urbanization, or socio-economic
level could not be determined. The less frequent use of toothpaste in the rural community
supports the notion that factors related to living conditions may affect the severity of
dental fluorosis.
No references or definitions are cited.
The study was well-conducted and had adequate study design. However, the study was
not designed for development of a dose response to fluoride as the emphasis was on
monitoring dental fluorosis in children and on attempting to explain other factors besides
fluoride in the drinking water that may contribute to fluorosis. Higher prevalence and
severity of fluorosis was expected in Arusha where the fluoride in the drinking water was
18-fold higher than in Kibosho. However, based on the study design, the prevalence of
dental fluorosis did not differ significantly between areas, but the severity was
significantly higher in Arusha (3 .6 mg F/l) according to TFI scores on tooth 2 1 . Apart
from fluoride in the drinking water, other sources of fluoride (magadi) may partly explain
the relatively high prevalence and severity of fluorosis in Kibosho (0.2 mg F/l).
Limitations of the study included:
o The amount and the fluoride content of consumed magadi could not be assessed.
o Water intake was estimated at the subject's present age.
o The duration of breast-feeding was not considered and the possible protective
influence of breast-feeding could not be assessed from these data.
Study design was not suitable for development of a NO AEL for fluorosis.
Study design was not suitable for development of a LOAEL for fluorosis.
Not suitable (X), Poor Q, Medium (_), Strong (_)
While the study was well-conducted, the study design was not conducive to provide data
for a dose-response. The study indicated similar prevalence of dental fluorosis in children
of both communities, but slightly higher severity of fluorosis in children living in the
community with greater fluoride concentration in the drinking water (Arusha). The study
did not address any issues of caries, plaque or gingivitis.
Prevalence and severity of dental fluorosis
19
January, 2008
-------
Bharati, P., A. Kubakaddi, M. Rao and R.K. Naik. 2005. Clinical symptoms of dental and skeletal fluorosis
in Gadag and Bagalkot districts of Karnataka. J. Hum. Ecol., 18(2): 105-107.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
Dental and skeletal fluorosis
Case control
India/ 6 villages in Gadag and 2 villages of Bagalkot District: 532 male and female subjects
surveyed from 6 villages in the Mundargi taluk (Gadag district) and 300 male and female
subjects surveyed from 2 villages in the Hungund taluk (Bagalkot district). Ten percent of
the households from each village were chosen for the study with at least one member of the
household exhibiting fluorosis. All members of the households chosen were part of the
study sample.
PROFILER'S NOTE: The ages or range of ages of the participants were not included in
the study report.
None described
Not described.
PROFILER'S NOTE: The profiler assumes since all members of the household were
included in the study that some of the participants (i.e. parents) had received long-term
exposures to the fluoride levels.
Only fluoride levels in drinking water were provided. Water in the Mundargi taluk ranged
from 4.0 to 10.5 ppm (Bharati and Meera Rao, 2001; Bharati, 1996) and water in the
Hungund taluk ranged from 2.04 to 3.2 ppm (Kubakaddi, 2001).
PROFILER'S NOTE: The applicability of this study for use in developing United States'
guidelines is limited as the values of fluoride exposure are much higher than those found
typically in the U.S. drinking water supply.
Participants were only assessed for the exposure to fluoride through the drinking water.
Analytical methods were not described. Only ranges for the fluoride level in the water were
provided; no other water parameters were measured.
The study was conducted in 6 villages of Mundargi taluk (Gadag district) and 2 villages of
Hungund taluk (Bagalkot district) in India that historically had fluoride levels ranging from
2.04 to 10.5 ppm fluoride. In each village, 10% of the households were selected with the
criteria for selection being that one person in the family was affected with fluorosis. A
checklist was developed using available literature and consultation with a nutritionist to
record the clinical symptoms of fluorosis. The symptoms were recorded by personally
interviewing each individual in the families chosen and by observations with the help of
local doctors. The symptoms were then tabulated and percentages calculated.
No parameters used for scoring either the dental or skeletal fluorosis were described. The
dental fluorosis was observed by examination (see Table 1) and the skeletal fluorosis by
clinical symptoms described by the participants (see Table 2).
No statistical methods were described.
Table 1 below is copied directly from Bharati et al. (2005). In Mundargi taluk, out of 532
participants, 328 (61.65%) had either dental fluorosis (25%), skeletal fluorosis (5.45%) or
both (3 1.20%). Among the 300 participants of Hungund taluk, 194 (64.67%) had either
dental fluorosis (35%), skeletal fluorosis (17%) or both (12.67%). In the Mundargi taluk,
20
January, 2008
-------
Skeletal fluorosis
browning of the teeth was the most common symptom of dental fluorosis followed by pain
and pus in teeth. Ninety five subjects had pitting and swelling and 86 participants had lost
their teeth. In Hungund taluk, lack of luster was the most common symptom followed by
browning of teeth with about 6 participants having lost their teeth. Overall, dental fluorosis
was more severe in Mundargi.
Titilt 1 Siraptimi uf dental fluoifi1;]* Jtnnn. clif fluiunbc -nlnnf ftum Mnudaigi and Hin;tmd taluk
i'r'V^t'L ^ui'^'i^i i £1 T^nr^ *,rt"_"
.Yiiif* Ft.mle T ttt
'>,„ rf.T^; H uft.nc' *Jimt.7-f B J \hiKt>7, Hw<~;.,»->
L,.u :.' tP] i-Si*"' ".r"-*-' 4:0" 4*t J£i •"'!.' '11I.--S4 iWn?:^-
'VJutp-n-h-f 4." 4\4i .hi iirts-i ii ".'-"'' "-i 1 _ QiiJJ-ij
B-u-niiPE'lteP-h l"i'M_> "1.5-C4I "•", 'i' 2" i j'i'4-& J^ifl "8> iHhifJci
P iifTi- «41uj :\:_J: - jt',5S"!i - .'i^iv
B i ^Tmigiuf Aim - 5,323' - 4,^ J>" " cl
Bir-nuif it 01111 >~ *j~"" l..C"i ?ev4-~ 21 1 2,1 aS i"«i ^"V T l
mil i
rSur.ui.] ^*e*epli _i * K - 4i2~- I'l if ^llt^ 11 '
L^ iftee'Ji 41 " " - 4\5' =2i 5 r ">•!' S'-Oft — i 5 3 'i
F i"p u nireirtli'1 rmJi ? pj-»i;eiirges
- Lj>jt c jj eo't'vi-^ 1 e t iiife elfroiatl.if1 Jij-1ii5<
PROFILER'S NOTE: The profiler agrees that the number of more severe findings were
observed in the higher fluoride area, Mundargi taluk; however, if the authors had provided
the data based on age groups and length of exposure, more useful information for
establishing a dose response would have been available for evaluation. Also, more details
in how the authors determined signs and symptoms are needed.
Table 1 below is copied directly from Bharati et al. (2005). For skeletal fluorosis, tingling
and numbness of extremities, back pain and bending were observed in a high number of
females in both areas. Males in both areas had more joint and knee pain. A higher
percentage of females were unable to walk properly or do normal work compared to males
in Hungund but the opposite was true in Mundargi taluk. Overall, skeletal fluorosis was
more severe in Mundargi taluk (the high-F communities).
Tahiti ^mjjtitm i-f "kdttiilfhwtii i tmong flnui uUc Tihiftt; bom \Iuudfii Ji and Hiingniidialul;
''^ t'.L"
'iA's rV w- Te\i>
'jnmjt'in ffi'n^vv' u'r.iiv'irj' Hnn^"»1 1 "11^71, ij, rTi, ,j>» f
Tijdug-wJUBilirj .1 '1 ' wi'oS jfil"!", PiluS«i --i!4'4 .filn-
i.£«r\T"iot "
J. . p ";ri 3" -!li ^^ ^I'r^.i •ij^'ii O-,TQ — .i|.>2>i
Bl'i. Jill "l.'l ""' y u -.« Miuv"-j H 2^ ~M )4!4rt"l; ,;[ "nil
IJi»e?irn "4i4n i1- J-iJl .5 ^"1^f5 <~\
"•liti'Ae niu il ~"i - 2 t t i *' _~i _i.f-i i^i^ i'!1 4i, *^'
'Se-ljt-j h •>! .-!!"._( '«i -tSi
P.m . uj'i S4-£' 2i. L? lirtii -* i i«_% h(3 J
-t'CUi i .' 1 -4i MIC "i4Si' i "•' JCi: >' .1 1 >~
;!lli' v If "ll IJUU 22)11 I"1 - -'l^"^! - 'i'~_
EPJ-I' '.c— '."> lit j: ,-*3°i _k:r?i jfi" C' *ii >
T c -i'ef i >aE picp-iS 1 a 2- _il ,. . if " i ^f,5( _> i~ li Qi4 i_
&'i'«te2' - - -.'."U ^1.^
Zau'cr mfl jwsrs I 1 \' I'll 6! - J !"~i iiu J_ «il
P:f til"- itm
'-ur-' z,:-. u-. 1'' •"! In) rJ
Ill^t vlU-S _ 1 . * - JllVil
. in* u >Vc 1 1 '" lui i
I'm-fe.u did 111 - in) ^ 2 ^'
"in'g=uiiT> eu"*" y 4 J "i 43 » i 13i
F>: "e ^piCenhsi mil tff-uer'g-
- LJ> iw' e" n i JH oft1 • 1 1 P t ]il»"c 1 a an t. a* ' uift ,.i
PROFILER'S NOTE: The profiler agrees that the number of more severe findings were
observed in the higher fluoride area, Mundargi taluk; however, giving the data based on age
groups and length of exposure would have provided more useful information in establishing
a dose response.
Also, more details in how the symptoms were determined are needed.
21
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/date
DFG/1-07
PROFILER'S ESTEM.
NOEL/NOAEL
PROFILER'S ESTEM.
LOEL/ LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
The people of Mundargi and Hungund taluk consuming water containing more than 2 ppm
of fluoride were suffering from both dental and skeletal fluorosis. Major symptoms of
dental fluorosis included lack of luster, browning, pain, pus and untimely loss of teeth.
Skeletal fluorotic symptoms observed included tingling and numbing of extremities, pain in
joints and knee, bending, stiff limbs, stiff vertebral column and unable to carry out the
routine duties. Preventative measures in these villages in the form of a supply of safe
drinking water and/or defluoridation of water is needed.
Bharati, P. 1996. Nutritional status and occurrence of fluorosis in selected villages of
Mundargi Taluk in Dharwad District. PhD. Thesis, University of Agricultural Sciences,
Dharwad.
Bharati, P. and Meera Rao. 200 1 . Epidemiology of fluorosis in Dharwad district. Journal of
Human Ecology. 14 (1): 37-42.
Kubakaddi, A.B. 2001. Epidemiology of fluorosis and educational intervention in Hungund
Taluk. M.H. Sc. Thesis, University of Agricultural Sciences, Dharwad.
PROFILER'S NOTE: The two references that are thesis publication are not likely to be
retrieved.
The study severely lacked details that could have been used for developing a dose response.
The ages of the participants including their length of exposure to the fluoride, actual
fluoride levels measured in the water (including analysis techniques), details on other
sources of fluoride, using a widely-accepted method for measuring the degree of fluorosis
and applying statistical techniques to the data were either not performed or not provided.
Application of the findings of this report to exposure conditions in the United States is
limited, as the levels of F concentration in US domestic drinking water are usually much
lower.
Despite the incomplete documentation and limited application of these findings to the US
domestic drinking water debate, this paper adds background information to the limited
dataset on skeletal fluorosis. No other sources of F, such as food or tea, etc., were reported
in Bharati etal (2005).
Focus of the study was on documenting the clinical signs of fluorosis. Water fluoride levels
for the individual households were not reported, and no evaluation was made of
confounding factors. Although the data did show that the community with lower fluoride
levels had fewer cases of severe fluorosis, the data are insufficient for a dose-response
analysis. Further, the populations studied are not comparable (regarding dental hygiene and
diet) to North American domestic water consumers.
The study is not suitable for developing a NOAEL for fluorosis.
The study is not suitable for developing a LOAEL for fluorosis.
Not suitable ( ), Poor (X), Medium ( ), Strong ( )
PROFILER'S NOTE: This study supports the hypothesis that the incidence of decayed and
missing teeth is increased when dental fluorosis is severe, especially in areas where access
to dental care is poor. There is a dramatic difference between the two populations for decay
and other severe dental problems.
Although this study lacks details and is incomplete, the results could possibly be combined
with more robust studies for weight-of-evidence that participants exposed to >2 pm
showed signs of dental and skeletal fluorosis, noting that a key piece of information missing
22
January, 2008
-------
CRITICAL EFFECT(S):
was length of exposure.
Dental and skeletal fluorosis
23
January, 2008
-------
Bottenburg, P., D. Declerck, W. Ghidey, K. Bogaerts, J. Vanobbergen and L. Martens. 2004. Prevalence and
determinants of enamel fluorosis in Flemish schoolchildren. Caries Research, 38:20-28
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
Dental fluorosis
Cross-sectional survey of dental fluorosis prevalence and severity
Belgium/Flanders: A representative cohort of Flemish schoolchildren (all born in 1989) and
selected from Dept. of Education records such that each child in Flanders had an equal
selection probability (stratified cluster sampling without replacement). Table 1 is copied
directly from Bottenburg et al. (2004) and provides the gender, age, number and geographic
location of the selected schoolchildren. For the final study, only data on 4, 1 28 children were
used. The study protocol was approved by the Ethics Committee (Catholic Univ. of Leuven)
in cooperation with the Dept. of Education and Flemish Regional Administration.
Table 1 . Sample characteristics of 1 1-year-old Flemish schoolchildren (n = 5,071)
Category Characteristic Children %
Gender male 2,668 52.6
female 2,403 47.4
pe°faPhl% . . Antwerp 1,440 28.4
localization (province)
Flemish Brabant 728 14.3
Limburg 871 17.2
Eastern Flanders 1,161 22.9
Western Flanders 871 17.2
Urbanization level city 549 10.8
town L230 24.3
suburb 954 18.8
countryside 2,338 46.1
Educational type private 3,414 67.3
state 685 13.5
province/municipal 972 19.2
PROFILER'S NOTE: The fluoride concentration associated with each individual province
was not included in the study.
none
Children were first examined at age 7 (1996) and then yearly with the final examination
taking place at age 12 (2001). Data in the study report was obtained when the children were
11 years old in 2000.
Children born in 1989 in five different provinces in Flanders (Northern part of Belgium) from
3 different levels of educational systems were chosen for this study. Although originally there
were 5,071 boys and girls, the data presented were based on 4,128 due to children moving,
illness or other unavoidable circumstances. All children were exposed to drinking water with
a fluoride concentration range of 0.47 to 1.47 mg/L.
Children were exposed to tap water which had fluoride concentrations ranging from 0.04 to
1 .47 mg F/L with a median value of 0. 16 mg F/L and a 75th percentile of 0.26 mg F/L. No
other water quality parameters were included. Children and/or their parents were questioned
on the use of fluoride supplements/toothpaste, tooth-brushing habits and diet.
24
January, 2008
-------
ANALYTICAL METHODS:
STUDY DESIGN:
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
Fluoride concentrations had been measured electrochemically in the water of different
municipalities since 1982. The long-term fluoride concentration in the municipal drinking
water was obtained from regional authorities (AMINAL, Brussels), water distribution
organizations (Vlaamse Watermaatschappij, Heverlee) or municipal authorities; analytical
techniques were not further characterized by the authors. The authors obliquely indicate that
local water supplies were not artificially fluoridated: "The presence of fluorosis in a region
with generally low natural fluoride concentrations in the drinking water and no artificial water
fluoridation could be confirmed in the present study."
Male and female schoolchildren from five different provinces within Flanders, Belgium were
used in the study. Data were collected on a total of 4, 128 children. See Table 1 under Study
Population for more details. The sample population was chosen using a technique of stratified
cluster sampling without replacement. Children were first examined at age 7 and then yearly
until age 12. Examinations were performed in a mobile dental clinic. Teeth were examined
for evidence of fluorosis and caries after teeth were dried using compressed air.
Questionnaires were given to the parents of the children and included the following
information: teeth brushing habits, use of fluoride toothpaste, use of fluoride supplements,
and dietary habits. Data for general health and urbanization level was supplied by the school
health records.
Fluorosis was recorded one time at subject age 1 1 (in 2000) and on the buccal surface of fully
erupted permanent teeth using the Thylstrup/Fejerskov index (TFI) (1978). (see Section 2 for
description)
Caries were scored at cavitation level in all deciduous and permanent teeth using a
WHO/CPiTN type E probe. Caries data were expressed as DMFT or DMFS in the permanent
dentition or draft and drafs in the primary teeth. On the day of the examination, the dentist
explained tooth brushing, provided dietary counselling, and distributed educational material.
Descriptive statistics were calculated to obtain frequency distributions of different variables.
Statistical analysis was also done on possible risk factors concerning prevalence and severity
of fluorosis. Univariable logistic regression analysis was performed to establish the effects of
the variables on the prevalence of fluorosis. Using "severity" as a continuous variable, an
analysis of variance (ANOVA) was performed using the same variables as independent
variables: 1) medical history, 2) sex of child, 3) level of urbanization, 4) tooth brushing
frequency, 5) age tooth brushing began, 6) quantity of toothpaste used, 7) use of fluoride-
containing dentrifice, 8) use of children's toothpaste, 9) use of systemic fluoride supplements,
10) age started supplements, 1 1) how supplements were given (in milk/not in milk), 12)
administration of fluoride supplements up to age 3,13) regularity of taking supplements, and
14) long-term mean fluoride concentration in the water system.
Non-parametric tests using Wilcoxon two-sample test statistics were performed to find
differences between groups of children. A simple logistic regression was used to establish a
relationship between fluorosis prevalence as explanatory variable and caries experience in the
deciduous and permanent dentition as outcome variable.
Chi-square tests were performed to establish relationships between time points of eruption for
teeth of the same quadrant (early = central incisors, and late = canines and premolars) and TFI
score to evaluate possible differences in fluorosis severity between these 2 groups of teeth.
Clinical examiners were pre-calibrated according to kappa values achieved after evaluation of
examiner scoring of projected slides illustration various degrees of fluorosis and caries. On
this basis, some candidate examiners were rejected from further participation.
In the study, most teeth were free of fluorosis and few had a TFI exceeding 3 ; three
individuals had a TFI of 5. Statistics indicated that brushing frequency, tap water fluoride
concentration, receiving fluoride supplements and taking fluoride supplements without milk
25
January, 2008
-------
all had a significant effect of the odds ratio for developing fluorosis. ANOVA identified no
variables significantly influencing severity. Figure 1 is copied directly from Bottenberg et al.
(2003). Chi-square test results were significant for severity differences between tooth 11 and
13 (p = 0.049), tooth 21 and 24 (p = 0.021) and tooth 11 and 14 (p = 0.005). Differences
between teeth 21 and 23 were not significant.
5SO-,
90-
80-
ro-
£ w-
— 50-
»S 40-
30-
M-
10-
0
On.
Oo_
On
On
Hnl
.
i-in.n.
nrra«.
Hn
On
toolh {FDI}
CZITFID cznTri 1 czuna ^BTFI 3 and Mgher
Fig. 1. Peivent.ige distribution oj t
scores tn tooth type (FDI tut-Jigit
and TFI Nuotbeis. in t»,n mtheseh letcr tu
the number of ftillv eiupted teeth available
tui exjiiunatiou
PROFILER'S NOTE: While fluorosis was observed, most subjects did not exhibit a TFI
score that was considered adverse. The article states that the teeth were grouped and presented
in Figure 1 above by the FDI two-digit system. In the FDI system, the first digit refers to the
quadrant (1 upper right, 2 upper left, 3 lower left and 4 lower right). The second digit refers
to the tooth type, i.e., 1 = central incisor, 2 = laterals, 3 = canines, 4 = first molars, up through
8 which are the third molars.
Dental caries
Table 6 was copied directly from Bottenberg et al. (2004). The author stated that both in the
primary and permanent dentition, decayed, missing, and filled teeth or surfaces were
significantly (pO.OOl) lower in children showing signs of fluorosis.
Table 6. Caries experience in the 11-year-old children in their deciduous or permanent dentition in (he group without
fluoKisis compa ted to the group with tluorosis
Variable, group
Range Lower Median Upper p value,
(minimum- quartile quartile Wikoxon
maximum) two-sample test
Deciduous dentition
dmf-l, no fluorosis
dinf-t fUiciKvB
dmt-s, no Buorosib
dmf-s, fluorosis
3,378
479
3,378
479
0-12
0-10
0-6 g
0-45
0,0004
<0.0001
Permanent dentition
DMF-T, no fltiore>sib
DMF-T. fiiiurosjs
DMF-S. no fluorosiN
DMF-SD, Chlorosis
UTS
4^v
U7S
0-11
0-5
0-26
0-25
0,0060
0.0067
Caries experience is tneti pti twtii idinl-t, DMF-T)or per surface (dmf-s, DMF-S),
PROFILER'S NOTE: Although the data show there were less decayed, missing and filled
26
January, 2008
-------
Other variables considered.
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date
REMARKS DFG
12/15/2006
teeth or surfaces, the TFI scores indicated fluorosis at a level that was not considered adverse,
as most values were below 3.
Table 5 was copied directly from Bottenburg et al. (2004) and shows the statistical
significance of all parameters measured. Tooth brushing frequency, fluoride supplement
usage, taking fluoride supplements without milk and tap water fluoride concentration (>0.7
mg/L) were significant risk factors when the presence of fluorosis on at least one tooth was
used as outcome variables.
Table 5. UnivaoaWe results ot tartar
influencing prevalence of fluorosis {child V™bte Odds ratio (95% C!t p value
saving ai least t tooth with TF! ^ I) as
outcome variable Sex (male vs. female) (Win 0.73-1.1 1) 0.32
Medieal history (contributory *s. nott (i.92t0.fi6-l,2S) 0.61
Uihimzmm 0.72
Trivuvv un 0.87(0.59-1.28)
SuNubvs uh 1.04(0.71-1.54)
loiintiwJevi utt 0.98(0.69-1.39)
Tuuth N ushing h.tbits
Rnis)un<>lro|iicno<£2x \« :x/day) 1.43(1.14-1.79) 0,002
UnsiipeniMrdattu Ueai^s upmvised ' 1.04(0.79-1.37) 0.79
StirtmgdijfottuMtli finishing* 2 waisvs.>2 years 0.83(0.61-1.13) 0.25
Quantity^) toothpaste 0.84
1 'pe,i-Mzt vs tullhtiisli 0.93(0.65-1.32)
Halttnushvi tullbiiish 0,93(0.73-1.19)
FluiitiJt tonthpj4c 0.34
Ustd in the past v soimmmirs us*. 0.52(0.21-1.29)
Ncvemseilvs lontffltttil.ltx 0.90(0.58-1.42)
L'ici't fkiKiidc-iedihidt ctnUieu t'l toothpaste 0.80
Inthepjstvs n>ntiniioiisusc 1.01(0.78-1.31)
Neveivs voutuiui'UMiw 0.88(0.58-1.32)
USE. ut s\stcn«t fluonde supplenif Eib
litres mur 1.3 1(1. 03- 2.68) 0.033
Suppkmui!;, suite 1,itiu 1 \eat if igevs. before 1.06(0.71-1.59) 0.76
1 akin run in milk it in milk l.«>(1.03-2.«8) 0.024
Vtlrmmstweilupto UiMi'niigevs longer 0.74(0.49-1.1!) 0,14
In^ul u idnnnistiatson v ^ leyuLiuiUmfflistratiaa I.08(O.E2-1.43) 0,59
1'at>»4t<.rlluuufk ioninwrHtno <().(I,)I
Bdun U U- alwi U " mg i 0.51 (0.38-0.69)
BrtAi*no.laudU"'\s ahoii it 7 mjU 0.58(0,41-0.82)
Fluorosis was present in about 10% of the children examined with most having a TFI of 1.
Tooth brushing frequency, fluoride supplement use and tap water fluoride concentration (>0.7
mg/L) were significant risk factors when the presence of fluorosis on at least one tooth was
used as outcome variables. Children having fluorosis had a lower risk of caries. Signs of
fluorosis are low in the prevalence and severity in Flemish children and correlated to a lower
caries experience. A higher proportion of teeth with TFI scores >2 was observed in later and
slower mineralizing teeth such as canines and premolars; nevertheless, the authors point out
that consideration of these latter data require precaution given the low number of late-
mineralizing teeth observed among subjects.
Bottenberg et al (2004) recommend that, if fluoride sources are to be eliminated in the future,
preference should be given to elimination of fluoride supplements (rather than decrease in
toothpaste use or brushing frequency).
The study did not correlate the incidence of fluorosis with a specific concentration of fluoride
and most of the schoolchildren exhibited fluorosis in the range that is not considered adverse.
The study adequately addressed all of the other variables that contribute to the fluoride
exposure of a child. This study did correlate reduced caries incidence with higher levels of
fluorosis, and also compared the degree of fluorosis severity in early-erupting and late-
erupting teeth.
January, 2008
-------
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM.
LOEL/ LOAEL
POTENTIAL
SUITABILITY FOR DOSE-
RESPONSE MODELING:
CRITICAL EFFECT(S):
Data are not suitable for estimating a NOAEL for dental fluorosis.
Data are not suitable for estimating a LOAEL for dental fluorosis.
Not suitable (X), Poor (_), Medium (J, Strong (J
The study is unsuitable to be used for dose-response modelling between fluoride in water and
degree of dental fluorosis; and, as presented, the data can not be used to ascertain the
threshold for severe fluorosis.
Dental fluorosis; the accumulative effect of low fluoride in the drinking water and other risk
factors resulted in no adverse findings in Flemish children.
28
January, 2008
-------
Brothwell, D. J. and H. Limeback. 1999. Fluorosis risk in grade 2 students residing in a rural area with widely
varying natural fluoride. Community Dent. Oral Epdemiology. 27: 130-6.
Brothwell, D. and H. Limeback. 2003. Breastfeeding is protective against dental fluorosis in a nonfluoridated
rural area of Ontario, Canada. J. Hum Lact, 19(4), p. 386-390
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
Dental fluorosis
Prevalence and Retrospective case-control study
Canada (Ontario): 1739 children in the 2nd grade (ages 7-8 years) from 55 out of 95 local schools
in a rural, non-fluoridated area of Ontario, Canada were involved in dental screenings. From
these, 1367 had erupted maxillary central incisors and were scored for fluorosis. The study
population consisted of children with a TSIF score of > 1 .
Of the 1367 children scored for fluorosis, only 752 (55%) of them returned with completed
questionnaires and samples of their home drinking water to be evaluated for fluoride
concentration. Of the 752, 175 had a TSIF score of > 1.
As explained above under Population Studied, the same children were screened to be used in the
control population. The criteria for the control population was a TSIF score of 0 and of the 752
children in the study, 577 had this score.
Based on the questionnaires, 3 9% of the children had lived in their current residence since birth
and 64.8% resided at their home since age 3 or less.
Children included in this study were part of a mandatory health program requirement for the
Wellington-Dufferin-Guelph Health Unit that provided dental disease surveillance. In the 55
schools chosen, 18 were high dental need schools, 30 moderate dental need schools and 7 low
dental need schools based on the number of children with an urgent need for dental treatment
from the previous year. Those children in the 2nd grade that were screened and had a TSIF score
of >1 made up the study group population and those with a TSIF score of 0 were used as the
controls.
Water samples from the children's homes were tested for fluoride concentrations.
Questionnaires included in the study inquired about breast feeding versus bottle feeding, use of
fluoride supplements, age at which tooth brushing was started, type of toothpaste used, amount
of toothpaste used, routine use of fluoridated mouthwash and any professional fluoride
applications.
Fluoride levels in the water were measured at the Faculty of Dentistry, University of Toronto,
using an Orion fluoride-specific electrode. The study states that the precision of the electrode in
the range of fluoride was within 1%. No other parameters were measured in the water.
1739 children in the 2nd grade (ages 7-8 years) from 55 out of 95 local schools in a rural, non-
fluoridated area of Ontario, Canada were involved in dental screenings. One examiner
performed the dental examinations in each school clinic using a portable light, mirror, explorer
and gauze. Central incisors were assessed for fluorosis by TSIF score.
Those children given a TSIF score were sent home with a questionnaire that assessed the
following areas for possible fluoride exposure: breast- vs. bottle-feeding, infant formula, skim
milk, fluoride supplements, toothbrushing, professional fluoride treatment and fluoridated
mouthwash. Also included were questions in regards to satisfaction of teeth appearance, years
residing in area, household income and education level. A sample of tap water was requested
and a sample vial provided.
Fluorosis was scored using the 8-point Tooth Surface Index of Fluorosis (TSIF) (Horowitz et al.
29
January, 2008
-------
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
Variables associated with
fluoride exposure
1984). The examiner kept laminated cards detailing the criteria and example pictures of each
grade during the examination.
Other possible fluoride exposures were identified using the questionnaire, as described under
Study Design, and a sample of tap water was analyzed for fluoride levels.
PROFILER'S NOTE: Details regarding the TSIF index are included in Section 2 of the report.
Data entry and analysis were done by the principal investigator using Epi-Info 5.0 and SPSS
for Windows 7.5. Bivariate analysis was used to identify possible important predictors for
prevalence and severity of fluorosis. Those indicating statistically significance (P<0.05) on
x2, t-test, or ANOVA were entered into multiple logistic regression to assess independent
effects. The relative effect of different variables was assessed by comparing resultant odds
ratios and 95% confidence intervals.
Intra-examiner reliability for the single examiner in assessing TSIF grades for the participants
was assessed by re-examining 55 students in one school 6 months after the initial
examination. A weighted kappa score of 0.75 indicated very good agreement. The 6-month
interval was used to lessen examiner recall.
Examiner training and calibration was done using sample photographs, concurrent in-school
examinations, and by developing a diagnostic decision tree. Prior to the initial examinations,
inter-examiner agreement when grading TSIF on 44 photographs was assessed. All
photographs were independently assigned a TSIF score by the examiner and two investigators.
A weighted kappa value of 0.89 showed excellent inter-examiner agreement in grading
fluorosis.
Table 1 is copied directly from Brothwell and Limeback (1999) and shows the TSIF scores for
the children in the study.
Table 1. Prevalence and severity ot cental fluorosis (TSif) bv «."\
TSIF scoro Prevalence [prouped TSiF s-ciw)
Sex Tot..: munK'i 01234 TS!F al (•'„) TSIF 2:2 CM
M 3o- 287 A? 11 3 0 79(21.6"..) 1J (3.S"..)
f 18" 3W 73 !.? 44 % (2-J.8".') 23 (h.0",0
Tola! ~2 377 I3fi 26 / -4 17? (25.3". -1 37 (4.9" i.)
PROFILER'S NOTE: The study indicated that the majority of the children had a TSIF score of
0 and for those with evidence of fluorosis, most were given a TSIF score of <2. According to the
NRC (2006), fluorosis is considered severe when the TSIF score is >5, indicating that none of
the children examined exhibited severe fluorosis.
Table 4 was copied directly from Brothwell and Limeback (1999) and showed the results of the
bivariate analysis. Bivariate analysis found that the fluoride exposures significantly (p<0.05)
associated with fluorosis (TSIF >2) were home-water fluoride concentration and fluoridated
mouthwash use. Those associated with a TSIF >1 were home water F concentration, breast
feeding length, fluoridated mouthwash use and professional fluoride treatments. Logistic
regression analysis for fluorosis (TSIF >1) showed the following had significant independent
effects: home water F concentration, duration of breast-feeding, use of fluoride supplements and
fluoridated mouthwash.
30
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
Table 4. Results of bivariate analysis
TS1F > I ("..) TS!P22{">.J
Home water floor ek*
•; 0 70 mg 1 22,3".. 4,8"..
--07Umg;i 31.3".. ns 18,?".-*
Breast- feeding duialion
- h month*; 27,2".- 7.9",.
b- 12 months l^i-T. 3.8'V
••12 months n 8". ' 1.7",, ns
•\j;e parent staitiL irushing child ?. t»>eth
• 1 year age 22 2'V 4,0",,
>3 vears age 6 ^"<>: 0,0"« ns
PiolesiMOnal fluondo treatment
No 5 !"„ 3,2°,,
\e> 243". ' 4,9",, ns
Fluoride Mjpplemwit use
• J year duration 26 7",, 0.0'V
1 via r to 2 vears. duration 11 ^",1 7,7''.,
"-2 \ears duration ^OU'n' 20,0'V ns
Fluoridated rnnuthwash use
\o 2271,, 4.t>"»
Yt'-i ?>(>.7"i' ns 13,3"«"
r- '0,05
; P-'O.OL
Figure 3 was copied directly from Brothwell and Limeback (2003) and shows a statistically
significant (p<0.05) difference in the amount of time children were breastfed and the rate of
fluorosis. The longer breastfeeding occurred, the less fluorosis was observed.
3°f
A 25'X
I 20'X
"1 is-^5
^ s^^^^ g y^^^^ |l ^§li$^' M
1 10"/'
5'^
at <#<$$&&$• §il$§liiii^ ^^^^fp'iil^if^i ^^^^ ^raS^k^'^'ir ^^^^^^ |^~
< t> 6lul2 >12
Ei Breast feeding duratMin (months)
"* « * » *.vj?*
Figure 3. Unadjusted fluorobis i ate by bieastfeeding; duration,
PROFILER'S NOTE: The profiler agrees with the findings.
The two journals reported on the same study but had two different objectives. The first article
(1999) was a pilot study and identified risk factors that accounted for fluoride exposures in an
area of non-fluoridated water. Brothwell and Limeback (1999) concluded 1) that fluorosis could
be a concern even in non-fluoridated areas, 2) that fluoride supplements should not be given
unless a home test for water fluoride level is performed and 3) that breast-feeding for more than
31
January, 2008
-------
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date:
REMARKS DFG/12-06
PROFILER'S ESTEM.
NOEL/NOAEL
PROFILER'S ESTEM.
LOEL/ LOAEL
POTENTIAL
SUITABILITY FOR DOSE-
RESPONSE MODELING:
CRITICAL EFFECT(S):
6 months may be beneficial to preventing fluorosis in permanent incisors.
In the second journal article regarding the same study, Brothwell and Limeback (2003)
concentrated primarily on the conclusion that breastfeeding for >6 months may protect children
from developing fluorosis in the permanent incisors.
None
The study adequately assessed most variables that contribute to a child's exposure to fluoride
and reported on which would be more statistically significant providing direction for future
studies. The degree of fluorosis observed in the report was not severe, however, and the profiler
is unsure whether the same trends observed would be similar in areas of high fluoride or severe
dental fluorosis. Adding the evidence of caries prevalence would have been helpful also to see if
any correlations occurred between the degree of fluorosis and the fluoride exposures. The
profiler agrees that breastfeeding longer than 6 months does appear to decrease the rate of
fluorosis.
This paper also includes information pertinent to relative source contribution analyses (Table 3
and others).
The data were not suitable for developing a NOAEL.
The data were not suitable for developing a LOAEL.
Not suitable (X), Poor ( ), Medium ( ), Strong ( )
While the study was not suitable for indicating a dose-response, useful information as to what
variables are important in assessing a child's total fluoride exposure was provided.
Variables involved in assessing total fluoride exposure and the effect on dental fluorosis in
young children.
32
January, 2008
-------
Budipramana, E.S., A. Hapsoro, E.S. Irmawati and S. Kuntari. 2002. Dental fluorosis and caries prevalence
in the fluorosis endemic area of Asembagus, Indonesia. International Journal of Pediatric Dentistry,
12:415-422.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL METHODS:
Dental fluorosis and caries
Cross-sectional survey in an endemic fluorosis area.
Indonesia/coastal East Java: Table 1 was copied directly from Budipramana et al. (2002) to
characterize the study population of children. Gender was not specified; subject children (N=
474) were between 6-12 years old and lifetime residents of the village. All of the villages in
the subdistrict of Asembagus, Indonesia had similar ethnic and socioeconomic status and
drinking water supplied by local wells.
Table 1, Grouping of tea villages is subdntrie! Aseaibf.Ein ;cto
dc.T,e-, i'ceord:n£ to Eueride content ;i: tLe dnnkir.,? wace:.
Flv.cnde
Yilkee Kan:e of content
group viliise :.ppcf; Mean SB n Si
A Sertcran 0 41~ 051 0-16(5 48 14!
Mojo s an 0 J6~ 48
Asecj 3s£iis C"-6fC 45
B Keduag Lull I; "33 CS! 49 97
Tii£ca;o 0 POC 48
C ?eraiite 1 01 T 2 2'j 0-684 47 142
G'.-.dana 2 317 41
Aivar-Aivar 2 41" 47
D Boati] 3 Cv5 5 16 0-195 4S 94
Wrinsm .inoru 3 2.3 46
PROFILER'S NOTE: A confounder exists as Budipramana et al. (2002) state that in 1990,
the local municipality attempted to supply water from another source to help reduce fluoride
exposure. The water was considered low-fluoride (0.45 mg F/L); this attempt did not work as
most of the inhabitants preferred to drink water from their own wells. The profiler is unsure if
means that some of the older children (i.e. 9-12 year olds) in this study would have been
exposed to this lower fluoride water (0.45 mg/L) for a period of time. The authors do not
clearly state if this source of low fluoride water was actually in the villages presented in the
study.
None
Subject children were 6-12 years old and were lifetime residents of the villages.
Levels of fluoride concentrations in a sample of local drinking water are included in Table 1
above in the Population Studied section. Data were also collected on each child's dietary and
residential history as well as drinking water consumption.
Children were assessed for exposure to fluoride in the drinking water as well as the main food
items in their diet. As most of the protein diet was fish or mussels, water from the local river
(the Banyu Putih River) and its fluoride concentration was (historically) measured.
A total of three samples of water were collected in plastic containers from each village ; one
from each of the three different wells. Fluoride concentration in the water was determined
using a Spectrophotometer UV-1201 (Shimatzu, Japan). No other water quality parameters
were included in the study.
33
January, 2008
-------
STUDY DESIGN:
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
PROFILER'S NOTE: The study states that the water samples were collected in October 1999
after a dry season and therefore, fluoride levels could have been higher than normal as the
levels tended to rise after a drought and lower during the heavy rain periods. It is stated that 3
samples of water were obtained from 3 different wells within each village but the authors do
not state if the samples were collected all at the same time.
Gender of the study population was not specified; subjects were between 6-12 years old and
lifetime residents of each study village. Subjects were chosen at random from 1 selected
primary school in each of the 10 villages studied. All of the study villages were located in the
subdistrict of Asembagus, Indonesia, and had similar ethnic and socioeconomic status.
Children were examined for fluorosis and the incidence of caries under natural light using a
sharp dental probe and mirror. Examinations for the study were performed one time during
October 1999. Children were also questioned about the main food items in their diet.
The DMFT index was determined according to the standards of WHO (1997). The degree of
fluorosis was determined using Dean' s index (WHO 1 997) as modified by Bischoff et al.
(1976). The degree of fluorosis was graded 0-3 depending on the extent of pitting and
mottling on the two teeth most affected; if the 2 teeth were unequally affected, the least
affected tooth was considered.
Four classes were derived:
1) normal teeth with no fluorosis. Scored as 0
2) teeth with white, opaque area with no brown staining. Classed as mild with a score of 1 .
3) teeth with extensive white, opaque, mottling, irregularly scattered, brown staining and
minute pitting. Classed as moderate with a score of 2
4) teeth with wide-spread brown staining and extensive pitting. Classed as severe with a score
of3.
The community fluorosis index (CFI) (Dean 1942) was computed by using the following
formula that calculates the average severity of scores.
CFI= £ (frequency x statistical weight)
No. of individuals
PROFILER'S NOTE: The profiler located the paper by Bischoff et al. (1976) and the authors
state they used their own method for determining the degree of fluorosis using Dean's index
as a guide. Bischoff et al. (1976) only uses scores of 0-3 (see above) and Dean uses scores of
0-4 (see Section 2).
Statistical analysis was conducted by using Kruskal-Wallis one-way ANOVA and multiple
regression analysis.
Table 3 is copied directly from Budipramana et al. (2002) and indicates the prevalence and
degree of fluorosis for children in each village. Overall, 96% of the children examined had
evidence of fluorosis, with most being graded as a 1 or 2, mild to moderate. There was a
statistically significant (reported as p=0.000) difference in CFI between the village groups
with the highest CFI being 2.03 in village group C (Perante, Gudang and Awar-Awar) which
had a mean fluoride water concentration of 2.25 ppm.
34
January, 2008
-------
Dental caries
Food and water sources
Table 3, Prevalence of fluorosis, degree of luorosis and community fluorosis index (= CFI) m subdistrict Asembagus.
Degree of fluorosis
village prevalence
group of fiuorosis 0123 CFI
A 92% 10 70 39 22 1-36
»=14I 7% 49% 27% 15%
B 100% 0 41 47 9 1-67
n = 97 0% 42% 48% 9%
C 9g% 4 23 85 28 2-03
pj=142 2% 17% 60% 19%
D 94% 5 28 38 23 1-80
n = 94 5% 29% 36% 24%
A+B+C+D <•«•'„ 1-71
4
PROFILER'S NOTE: From the data provided, the degree of fluorosis did not increase as the
fluoride content in the water increased and the majority of cases were moderate (2).
Table 2 was copied directly from Budipramana et al. (2002) and shows the prevalence and
distribution of caries in primary and permanent teeth. Reported means indicate that 62% of
the children were caries-free for permanent teeth and 67% were caries-free for primary teeth.
In this study, there were very few teeth diagnosed as missing or filled, so only data for
diseased teeth (DT- permanent or Dt- primary) were discussed. There were no statistically
significant differences in the number of diseased teeth between groups of villages in either the
permanent or primary teeth, but the number of diseased permanent teeth tended to increase as
the fluoride levels increased. The number of diseased primary teeth decreased as the fluoride
increased.
Table 2. Distribution of DT, dj, prevalence of caries in permanent and primary teeth of clirldren in sufedistnct Asembagus.
Mean teeth/person Percentage canes free
group Permanent Primary DT dT Permanent Primary
A 15-621 8-120 0-461 0-893 62% 64%
B 14-748 8-864 0-556 0-876 67% 56%
C 16-514 7-460 0-662 0-866 62% 69%
D 14-896 8-611 1-031 0-329 55% 82%
4
PROFILER'S NOTE: The study did not correlate the degree of fluorosis with the incidence of
caries, only the incidence of caries with the fluoride concentration of the water.
Dietary percentages offish, salted fish and mussels consumed as main (protein) food items as
well as drinking water consumed from individual wells for each village group were included
in Table 4 of the study report. The fluoride content of the Banyu Putih River was reported to
be 5.0 to 5.2 mg F/L (Rai 1980); this river passed by the study villages and was the source of
locally consumed fish, etc. The river is also used as a source of irrigation water for crops.
The authors further note that individual wells located near the river exhibited a high fluoride
content (2.07-3.25 mg F/L), while wells located at a greater distance from the river in a less
heavily irrigated area exhibited relatively low F (0.41-0.90 mg F/L).
Table -1, Sc-urte of drjiki&e water and food of tie village groups
of siiijdistr;;* AsembiEu;.
Percentage of
v.-atei from Mam food coLiUarumt
Village i::d;vidiHl
group \ve".l& Fi;h Salted fbh Mustels
A S53= 69°= "Dco ~""c-
B ~9°c PC>% 5i% 573c
C': ~1^, •" j£3 £Q® ^£3
. fi &J $ Ofo i JDo
D 3S5= SSS 5"% 60%
PROFILER'S NOTE: The percentages stated in Table 4 (Budipramana et al. 2002) of the
main food items do not correlate to the percentages stated in the summary at the beginning of
the paper. The profiler is unsure why there is a discrepancy.
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
DFG/12-06
PROFILER'S ESTIM.
REVIEWER'S NOTE: The Budipramana et al (2002) paper does not include a current
measurement of the Banyu Putih River fluoride concentration, which is expected to fluctuate
with time of year (drought, rainy season, etc.). The reported river water concentrations
published by Rai (1980) were collected approximately 20 years before the time of the
Budiprama et al (2002) study and may not be representative of conditions at the time of the
Budipramana et al (2002) study.
The principal findings from this study provided evidence that dietary sources other than
drinking water should be taken into account in studies of fluorosis. There was no relationship
between fluoride levels in drinking water and caries, despite significant differences in dental
fluorosis. The study also indicated that dental caries in primary teeth of children exposed to
high concentrations of drinking water fluoride was lower than for those living in localities
with less fluoride. There was a positive correlation between CFI and fluoride concentration
of drinking water.
PROFILER'S NOTE: In the study summary at the beginning of the paper, the author states
that caries prevalence in the subdistrict was 62% for permanent teeth and 68% for primary
teeth when actually the data in Table 2 (Budipramana et al. 2002) indicates that 62 and 67%
were the average percentages that were caries free in the permanent and primary teeth,
respectively.
Bischoff, JL, EHM Van Der Merwe, DH Relief, FH Barbakow and PE Cleaton Jones. 1976.
Relationship between fluoride concentration in enamel, DMF index and degree of fluorosis in
a community residing in an area with a high level of fluoride. Journal of Dental Research.
1:3742.
Rai, IGN. 1980. The relation between prevalence of endemic hypoplasia teeth in children
with fluoride content in the drinking water, urine and carious teeth. PhD thesis. Surabaya,
Indonesia: Airlangga University, 62-84.
World Health Organization (WHO). 1997. Oral health survey, basic methods. 4th edn,
Geneva: WHO: 35-35, 41-6.
The profiler finds this study unacceptable due to a confounder identified and data that are
stated incorrectly. While it is possible that the study summary was added at a later date and/or
there was a problem with translation, the data do not reflect what is stated in the body of the
report. Therefore, a dose-response could not be derived. The study also did not correlate the
degree of fluorosis with the incidence of caries, only the incidence of caries with the fluoride
concentration of the water.
Noted fluctuations in the F concentrations of individual drinking water wells due to local
climate variations (drought vs rainy conditions) and proximity to irrigation channels
containing river water generate uncertainty when attempting to compare the amount of F to
which subjects' teeth were exposed during enamel production and tooth eruption. The authors
are aware of this and point out that F concentrations in well water fluctuate, with high
concentrations during drought and lower concentrations after heavy rains; and that there is
"large variation" in F concentrations between individual wells within a single community. It
is thus doubtful that the reported well-water fluoride concentrations per village are
representative. The above is further evidence that the Budipraman et al (2002) data are likely
compromised.
The reviewer concurs with the profiler's findings.
Profile is approved for project officer review.
Not suitable for development of a NOAEL.
36
January, 2008
-------
NOEL/NOAEL
PROFILER'S ESTIM.
LOEL/ LOAEL
POTENTIAL
SUITABILITY FOR DOSE-
RESPONSE MODELING:
CRITICAL EFFECT(S):
Not suitable for development of a LOAEL.
Not suitable (X), Poor ( ), Medium ( ), Strong ( )
Dental fluorosis and caries
37
January, 2008
-------
Chen, B.C. 1989. Epidemiological study on dental fluorosis and dental caries prevalence in
communities with negligible, optimal, and above-optimal fluoride concentrations in
drinking water supplies. Zhonghua Ya Yi Xue Hui Za Zhi. 8(3): 117-127.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
Dental fluorosis and dental caries
Cross-sectional survey (conducted August, 1987 to July, 1988) of dental caries and dental
fluorosis.
Taiwan, Shenkang Hsiang Province: children (2,669 boys and 1438 girls) aged 6 to 16 yr
old and lifelong residents of 14 communities.
None
6 to 16 yrs, and beginning inl 971-72 (for 16 yr olds), and beginning in!981-82 (for 6 yr
olds).
Fourteen study sites were divided into six exposure categories based on water fluoride
levels; negligible, optimal, 2 x optimal, 4 x optimal, 5 x optimal and 7 x optimal.
Negligible fluoride levels were considered to be <0.4 ppm and optimal levels >0.4<0.7
ppm. The study author notes that the great majority of the population acquire their drinking
water from shallow wells, and the rest from deep-wells. However, it was reported that
beginning in June 1981 a communal water supply (non-fluoridated with a fluoride level of
0.09 ppm) was available to the population and was supplied to all the schools in the
province. Other exposure factors, such as dietary contributions to fluoride intake and the
use of fluoride dentifrice and supplements were not evaluated.
Fluoride levels were determined with an Orion SA 270 Ion Specific Electrode. Well water
from 98 sampling sites was collected during the four seasons, and the annual average of
four readings from each location was determined. Other water quality parameters, such as
calcium levels, were not reported for any of the water supplies.
Dental fluorosis and caries incidence were evaluated in children 6 to 16 yrs old (2,669 boys
and 1438 girls) residing in 14 communities in Shenkang Hsiang province, Taiwan. The
communities were divided into six exposure categories based on water fluoride levels as
defined by Chen et al (1989); negligible, optimal, 2 x optimal, 4 x optimal, 5 x optimal and
7 x optimal. Negligible fluoride levels were considered to be <0.4 ppm and optimal levels
X).4<0.7 ppm. The recommended range of optimal water fluoride concentrations used in
the study was 0.4-0.5 ppm for the tropical zones of Taiwan and 0.6-0.7 ppm for the
subtropical zone, and is based on zone-specific water consumption rates. Dental fluorosis
was scored using TSIF and the Dean Fluorosis Index. Caries incidence was scored using
the DMFS index. Statistical analysis was carried out using Student's t-test, Chi-square,
analysis of variance, and multiple regression.
Tooth surface index of fluorosis (TSIF) and the Dean Fluorosis Index (FI) were utilized in
the examinations (see Section 2 for definitions and descriptions). Radiographs were not
taken. Caries incidence was scored using the DMFT index (see List of Acronyms for
definition). A Community Fluorosis Index (CFI) and the public health significance of the
fluorosis problem were determined for each village using the methodology of Dean
(1946)..
Student's t-test, Chi-square, analysis of variance, and multiple regression. The level of
significance used was p<0.05. Scheffe's method for multiple comparison was used to
compare pairs of CFI scores and pairs of DMFT scores. No other specific information was
provided on statistical methods.
38
January, 2008
-------
RESULTS:
Results of the study are shown in Tables 1, 2 and 5 copied directly from Chen (1989):
Table 1. The relatiomlnp 01 d«nl.»l writs, denul fiootosu, !K twj l'i Mem nuor*- ,0 ,,,,„ ^
,,,,,, ,nnM 3"eUk*'by v-nMH l1cntf f|uirws (ppmj eoii; -.ti. u
Kt'uu'i "'i .' "i 2*>J -d ~l "> >' '" ^>'!?Jt« 0,21 V.'-K!*
BeiehLirn - ,• ' ) 9? 3.5 8.3 OK' V-j.'ne 0,22 i\. ife. •!«
liu'-nh ui i, '',!-) 487 2.9 3,7 1 .OS ViV, '»« 0,25 \>>li> W?
CT i ' r«h 327 2.5 6,1 i'. - NJ,>r.r»n 0,43 , pt,<> „
CH,UTV >i • '-.i •-> 349 2,3 4-8 h.^ V^ ive 0.4S < C'n>«
S'uiK-'^it '. i 98* 2,6 7.2 'i<+ N.\>I*HJ 0.48 n*>'i,i
f^rai.','* 265 2.3 22.6 t> 37 V.jan 0.98 2 \ uptm.i
L'JufsHn^t i'rt "'! 420 1.4 39.1 0 ". Sl«>-'i 2,40 4x.i>i0.4<0.7
ppm.
Information on the relationship between fluoride levels, fluorosis scores, and CFI are given
in Table 1 copied directly from Chen (1989):
Table 2. Percentage oiiCnbutkin of chi'Mier accurdJing to Dean's fluorosis score xA water fluor.de I*v«l, with
commtittity t1uorfi»t& ihd^x ^o'W.H
W0te< tlw.iat \o .), , Co I1..U.1KV
• , , L'^a s!luu out. Kore ii.u....nj
U'»v, c i.ti.wn f.^.j.i-.^
* U.5 1 254 Iiuliv
Ne,!lipihle HS1 ^9 " '.', 4.9 V- 41"- !„, , - - .i.t«
OpUf.J Ifitit) s&,2 't *,J<: 36"- •>"""- (i,ll> 1..3
2 x uDTii.131 S4y 6"* " " 12.2 '"• " 1 '4 '."• *>&** OK", n 2 ',', n.34
4.\jp:ijtul 42 i 4J> 1 ."':'"• 86^ (> J '" A.J'f , 0 "2
SvoriiTii <»!2 :","', 0.6 '? :»!•: 316" _\? '; ii.j" i.'is
7.\jpcp'd 3K< IZ.i 'o T.6«' i«2J', »82" 12,fi ',' , ? '- ; 61
A dash ( — ) equals 0%; Fluorosis scores: 0 = none, 0.5 = questionable; 1 = very mild; 1=
mild; 3= moderate; 4 = severe.
The report indicates that compared with the negligible exposure group, DMFT scores were
10.7% lower in the optimal fluoride area, 14.3% lower in the 2 x optimal area, 50.0%
lower in the 4 x optimal area, 42.9% lower in the 5 x optimal area, and 42.9% lower in the
1 x optimal area (statistically significant, p<0.05, only in the latter three groups).
Table 5 from Chen (1989) indicates that older children exhibit higher levels of fluorosis
than younger children as indicated by the Tooth Surface Index of Fluorosis (TSIF, see
Section 2); TSIF scores of 1-3 indicate parchment-white areas of enamel of gradually
increasing size but no pitting or brown stain; a TSIF of 3 indicates white discoloration over
>2/3 of the enamel surface). Chen et al do not speculate on the observed finding that
"tooth surfaces of 13-16 year olds are more affected by fluorosis. . .irrespective of water
fluoride level" (p. 122).
39
January, 2008
-------
t'cr&'iK^e dt;lr*nt»ra of T.Slt sum" »o» all peimaiiti.I flM it:iL*rv H £\>™n-'nrti« wi;t>.
JtiT flwiriJa U'MfU jiojlO'Tj H« *JJB ,ituuti
Kn.cn -f Li. I? .Mi-r •>! "t 'ill'" scares
*''•"•'' !' -IJ
•HI 16 4 ! 7 f.'>? -''.4.1
H NO i;, H :<5.:>l .1 f^
• f. o vi >1) are similar (i.e., in the
Galagan and Lamson (1953) studies, 3-10% of the population in communities in the
southwest US with 0.4-0.5 ppm fluoride in their water had fluorosis, compared to 4.6-7.2%
reported in the Chen study.
Chen reported that children 12-13 yrs old had higher levels of fluorosis than 8-9 yr olds.
40
January, 2008
-------
PROFILER'S ESTEM. NOAEL
PROFILER'S ESTEM.
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
Dean (1942) states that 12-14 yr olds were the preferred age group for fluorosis studies in
areas with low fluoride water levels because "this permits the examination of a group in
whom a high percentage (approximately 94%) of the permanent teeth have erupted". Dean
further states that inclusion of children as young as 9 years old "seemingly makes little
material difference when the group has been exposed to relatively high water fluoride
levels (e.g., over 2.0 ppm of F)", but would introduce an "error of considerable magnitude"
if included in studies where fluoride levels were lower, because fluorosis in non-erupted
teeth would not be included in the analysis.
Could be possible with further statistical assessment.
A low incidence of severe fluorosis was found in the 0.8-1.4 ppmF (2X optimal), 2-3.5
ppm F (5X) and 2.8-4.9 ppm F (7X) groups, but not in the 1.6-2.8 ppm F (4X) group.
Therefore the data need to be examined statistically to determine whether the incidence in
the 2X group is significant.
Not suitable Q, Poor (X), Medium (_), Strong (_)
Severe fluorosis (DPI of 4) was reported in three of the six exposure groups (see Table 2
above); however, a dose response was not seen across the 2X and 4X optimal groups;
therefore, modeling may not provide a statistically reliable result.
41
January, 2008
-------
Chibole, Opati. 1987. Epidemiology of dental fluorosis in Kenya. J.R.S.H., p. 242-243
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
Cross-sectional survey of native inhabitants within specific provinces/districts of Kenya,
Africa. There were eight provinces evaluated and they included: Central, Coast,
Eastern, North Eastern, Nyanza, Rift Valley, Western and Nairobi. The provinces were
further divided into districts (see Table 2 below in Results section).
Over 34,000 people (ages or gender not specified) native (born and raised) to each
specific region were examined.
none
Individuals were examined from 1979-1982 to determine the prevalence of dental
fluorosis.
Individuals from each specific Kenyan region were examined separately for evidence of
dental fluorosis occurring from exposure to the ground water. There were eight (8)
different provinces in the study with a further breakdown to districts within each
province (See Tables 1 and 2 below under Results section).
Study participants underwent a mouth examination with dental probes and mirrors using
natural light as the source of illumination.
Analytical methods for measuring fluoride concentrations in the water were not
provided in this study.
Over 34,000 people from different areas in Kenya were examined at various schools,
hospitals and the Dental School of the University of Nairobi with mouth mirrors and
dental probes using natural daylight. These surveys took place during 1979-1982 and
included only those native to the regions listed. Individuals were examined for evidence
of dental fluorosis.
The following scale was developed by Chibole (1987) to be used to assess the degree of
fluorosis; however, no data on the prevalence of study participants exhibiting each score
were reported.
Degree 0: no mottling
Degree 1 : white or light brown patches on parts of the coronal surfaces of teeth
Degree 2: dark brown patches on parts of the coronal surfaces of teeth
Degree 3 : brown discolouration of the entire crown
Degree 4: brown discolouration associated with fracture of the enamel and presence of
pitting or cracks on enamel.
Chibole (1987) stated that data were used from an earlier report (Gitonga and Nair,
1982) that compared the population with fluorosis to the proportion of those ingesting
borehole water having fluoride levels greater than 1 ppm. A positive correlation was
made with the Pearson's "r" value being 0.65, and the probability that this correlation
occurred by chance was less than once in a hundred (pO.Ol). However, no data
specifically characterizing the water fluoride levels by province or district were
provided.
Results for the incidence of fluorosis are provided in Tables 1 and 2 copied directly
from Chibole (1987). The author did not provide data characterizing the observed
fluorosis prevalence by score (i.e. Degrees 0-4). Table 1 shows the % fluorosis present
in each province with the total incidence of fluorosis in Kenya being 32.2%. Table 2
42
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
indicates a further breakdown of data by providing the incidence of the population with
fluorosis from the districts within each province.
SSta : Pwateieeof fcwosis b^r Wlth fluorot
*3Bllnin 1 or some other criteria.
Although the national level for the prevalence of dental fluorosis in Kenya is about
32%, the prevalence was observed to be higher in areas where the population is totally
dependant on ground water (i.e. Rift Valley; see Table 2 of report). The author states the
need for more frequent monitoring of the ground water for fluoride levels and for
effective methods of defluoridation of water to stop endemic cases of fluorosis.
Gitonga, J.N. and K.R. Nair. 1982. The rural water fluorides project technical report.
43
January, 2008
-------
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
DFG/11-06
and
12/14/06
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
IDRC/University of Nairobi and Ministry of Water Development, Kenya.
The study is not suitable for development of dose response for fluoride. Although the
author provides an assessment on how the degree of fluorosis was evaluated, neither
these data nor results were included. The degree of fluorosis present in the population
examined would have been extremely helpful for evaluation of this study. Further,
actual measured values for the groundwater fluoride concentration from a specific
province were not included. As a consequence the profiler is unable to determine a
correlation between fluoride exposure levels and the incidence of fluorosis. The data
showed wide variation in the incidence of fluorosis between districts within the same
Province. The profiler is unsure if other sources of fluoride could be causing these
variations because no fluoride concentration levels were included.
The profiler is unable to derive any values from the study due to insufficient data
reported.
The profiler is unable to derive any values from the study due to insufficient data
reported.
Not suitable ( X), Poor ( ), Medium ( ), Strong ( )
Dental fluorosis
44
January, 2008
-------
Cochran, J.A., C.E. Ketley, I.E. Arnadottir, B. Fernandes, H. Koletsi-Kounari, A-M. Oila, C. van Loveren,
H.P. Whelton and D.M. O'Mullane. 2004a. A comparison of the prevalence of fluorosis in 8-year-old
children from seven European study sites using a standardized methodology. Community Dent. Oral
Epidemiol, 32 (Suppl. 1): 28-33.
Cochran, J.A., C.E. Ketley, L. Sanches, E. Mamai-Homata, A-M. Oila, I.E. Arnadottir, C. van Loveren,
H.P. Whelton and D.M. O'Mullane. 2004b. A standardized photographic method for evaluating enamel
opacities including fluorosis. Community Dent. Oral Epidemiol, 32 (Suppl. 1): 19-27.
PROFILER'S NOTE: Both articles were published in the same journal and will be evaluated in the same
profile. The 2004a article presents the study details and the 2004b article more thoroughly describes the
photographic method used for standardizing study data.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL
POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL
METHODS:
Dental fluorosis (enamel opacities)
Cross-sectional survey
Europe: Cork, Ireland; Knowsley, a suburb of Liverpool, England; a suburb (name not given)
in Athens, Greece; Haarlem, a town near Amsterdam in the Netherlands; Oulu in the north of
Finland; Reykjavik, capital of Iceland; and Armada and Setubal, both suburbs of Lisbon in
Portugal. Approximately 300 eight-year-old schoolchildren were randomly selected from each
of these seven European sites.
None
Fluoride levels from water sources were obtained for a 10-year period (1988-1998) including
the 8 years the children were exposed. The children were asked about their residential history
but the residential data were not included.
Tl
P£
a
ot
le following table shows the concentration (ppm) of fluoride in the water and the number of
irticipants at each specific site. Only the water in Cork, Ireland was artificially fluoridated (to
concentration between 0.8 and 1.0 ppm); fluoride in drinking water occurred naturally at the
tier sites.
Site
Cork (Ireland)
Knowsley (England)*
Oulu (Finland)*
Athens (Greece)*
Reykjavik (Iceland)*
Haarlem (Netherlands)**
Almada/ Setubal (Portugal)*
No. of children
324
315
314
287
298
303
210
*Fluoride concentrations reported annually.
**Fluoride levels fluctuated between 0. 1 and 0. 15
Fluoride concentration
(ppm)
1.0
<0.1
O.01
O.01
0.05
0.13
0.08
ppm for the 10-year period.
Only fluoride in the tap water from each area was measured. Parents of the participants were
given a questionnaire to inquire about use of fluoride supplements, history of living in area,
age when toothpaste was first used, and the amount and type of toothpaste used. Care was
taken to ensure the same concepts were expressed in the different languages on all of the
questionnaires.
Fluoride levels in the local drinking water were obtained for 10 years, 1988-1998, including
the period the study children were exposed. How the water was analyzed for fluoride was not
included in the study report. No other water parameters were reported.
45
January, 2008
-------
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL
METHODS:
RESULTS:
Dental fluorosis
Data were collected from Fall 1997 to February 1998. Several schools in each area
representing a wide socioeconomic basis were chosen. From these schools, children in the 8-
year-old range were invited to participate. Cochran, the study lead author, was responsible for
teaching one dentist in each area the photographic technique that was used in the study. All
seven dentists used the same camera with an identical set-up (i.e. lenses/flashes). Film used
was bought at the same location and originated from the same production batch.
The study author had several "training" sessions with dentists from each location to ensure the
photographic technique was similar. This involved a training session at Cork and also two
"pilot" studies at each location. After each pilot session, discussions were held and technical
problems addressed and corrected prior to the actual study.
Before being photographed, each child cleaned their teeth without water or toothpaste using a
new toothbrush. For the photograph, each child was positioned against a wall, a cheek
retractor placed and the incisors kept closed with edge-to-edge contact. A piece of damp
cotton wool was used to keep teeth damp. Two timed photographs were taken of the
permanent, maxillary incisors, one after 8 seconds for a wet photograph and one after 105
seconds for a dry photograph. While taking the photographs, the camera was held at a 45°
angle to minimize reflections and lip shadows. All film was then sent to London and
developed in one laboratory. The resulting transparencies were then randomized, viewed
blindly and scored for fluorosis by Cochran.
Questionnaires were also completed to determine if fluoride supplements had been used,
residential history, age when started using toothpaste and the type and amount of toothpaste
used.
From the total of 5250 transparencies taken, 114 (2.2%) were not suitable for analysis. During
the study, the intra-examiner reproducibility of each photographer was measured by taking a
repeat photo for 15% of the transparencies; inter-examiner reproducibility was measured by
the study author visiting each study site and repeating 15% of the photographs.
The photographs taken after 8 seconds, when the teeth were still wet, were examined for
fluorosis based on the Developmental Defects of Enamel (DDE) index (FDI Commission on
Oral Health 1982), and those taken after 105 seconds, when teeth were dry, used the Thylstrup
and Fejerskov (TF) index (Thylstrup and Fejerskov 1978).
PROFILER'S NOTE: The DDE index (FDI Commission on Oral Health 1982) used in the
study is not recognized in the NRC document (2006) as a true measure of fluorosis as it
emphasizes aesthetic concerns and is not based on etiologic considerations. The TF index is
described in Section 2 of the report.
Logistic regression was used to determine which variables to which the children were exposed
were most important in relationship to fluorosis. To summarize the variability in the data for
the range of time periods when the photographs were taken, box-and-whisker charts were
plotted. The level of agreement between pairs of grades was determined using Cohen' s kappa
statistic (Cohen 1960).
Tables 1 and 2 are copied directly from the study report (Cochran et al. 2004a) and show the
prevalence of fluorosis in both wet and dry teeth. With the DDE index, Cork had the greatest
percentage of diffuse lesions with Haarlem having the greatest overall prevalence. In the TF
index, Cork and Haalem had the greatest number of children (4) with the most severe TF score
(>3); although most children exhibited only mild fluorosis.
Table 1. Percentage of subjects with at least one permanent maxillary central incisor affected
by diffuse or demarcated opacities or hypoplasia and overall prevalence of enamel defects as
graded by one examiner from transparencies taken of "wet" teeth measured using the DDE
index.
46
January, 2008
-------
Reproducibility for the
DDE and TF indices
Variables in fluoride
exposure
I in wain Pidu-t Demarcated
*iilc mp'-pi n c,\ (%) Hypoplasia(%)
1 nil t It 324 M 16 1
KnimiV) J'l 315 12 19 2
Uu'u IMS1 314 IS 18 2
\thins ,».!! 287 2s 18 0
Kivkwuk .H« 298 "C 24 2
Rindem !• H ,303 r-r> 25 3
Mm.iJ.1 He:dbal <)>*- 210 31 20 1
Table 1. Percentage of subjects with each grade of the TF index (highest grade of 1 teeth) as
graded by one examiner from transparencies taken of 'dry' teeth.
1 in water TF gra.de (highest of two teeth; %)
Si to tppmi n Q 1 2 >3
Cud II! 32r- 11 59 26 4
Knmv-'ev - ai 314 31 54 11 1
CXil-a ' «IU>1 ir* IS 61 21 0
AlhiTb, -,nni
ph'l'vnplvr ^^^rv'jhms ind Hppi indkippi in.1 I ippi
b us' -.'. i'»i '. .
!•> >,, t,l> } IV '\.-iy i.ndi ' -.1 i 'll>
t. (18) ~V7 »)'", <;" ,
i-Ciiiiid" iin'i,«,i' i"°iv.inji
D CMS '• '1 ""
i'.,J u,»,,Ji i *' i, «il' 1 , 'M d.lil-i
E 06) li Iim 1
11 \-rv;iid> 1 **i I'.TV- c i Ji !•*! ',»rvmJ'
F 08) "> * • * 1 ' "'*'
>M V ry , ili '!,",< »ili il-^Hip
G 08) , , U't',
'if ,M ,1 ill i 1 m A.ry r«iJ> '"• «. fiJ'
The variables found to be the most statistically important were artificial water fluoridation
(odds ratio of 3.53 (95% CI of 2.52 to 4.93) and the prolonged use (> 2 years) of fluoride
tablets (odds ratio of 2.17, with 95% CI of 1.60 to 2.95). Cork had the most diffuse opacities
due to the artificially fluoridated water; however, Haarlem (0.13 ppmF) andOulu(<0.01 ppm
F) also had comparable numbers to Cork. When investigating the use of fluoride supplements,
no children had used them in Cork, but 44 % of the children had fluoride supplementation for
47
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE
NOT FOUND IN NRC
(2006)
PROFILER
'S
REMARKS
Initials/date:
DFG/1-07
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM.
LOEL/ LOAEL
POTENTIAL
SUITABILITY FOR
DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
> 1 years in Haarlam and 58% had in Oulu. In the report, children exposed to water fluoride
levels of 0.8 to 1.0 ppm were 3.53 times more likely to have a TF index grade 2 or more and
those exposed to fluoride supplements for > 2 years were 2 times more likely to have enamel
fluorosis.
PROFILER'S NOTE: The amount of fluoride in the supplementations used were not
reported.
In the article presenting the data on the study (Cochran et al. 2004a), the report found the
prevalence of fluorosis to be highest in Cork, Ireland (artificially fluoridated water supply).
Prolonged use of fluoridated supplements (tablets) was also associated with a significantly
increased risk of dental fluorosis. For the second article describing in detail the photographic
technique (Cochran et al. 2004b), the conclusion was the photographic method was robust and
reproducible when used in seven European sites.
Cohen, J.A. 1960. A coefficient of agreement for nominal scales. Educational Psychol.
Measurement, 20:37-46.
While the study adequately convinces the profiler that the photographic method for assessing
fluorosis could be a viable method to collect and compare data from several regions, the study
was not designed to be used as a dose-response study. Some deficiencies were the use of a
DDE index that the NRC (2006) does not endorse as a valid method for assessing fluorosis.
Also, most of the children examined had TF scores of <3 which indicates onlya mild
fluorosis. The data indicated similar results in the TF scores and fluorosis prevalence in the
area of highest community water fluoridation (Cork, 1 .0 ppm F) and some of the lower water
fluoride concentrations such as Haarlem (0.1 3 ppm) and Oulu (O.Olppm). When
investigated, it was found that use of fluoride supplements in Haarlem and Oulu had
contributed to the observed dental fluorosis.
Data were not suitable for developing a NOAEL.
Data were not suitable for developing a LOAEL.
Not suitable (X), Poor ( ), Medium ( ), Strong ( )
Dental fluorosis (enamel opacities).
48
January, 2008
-------
Cortes, D.F., R.P. Ellwood, D.M. O'Mullane and J.R. de Magalhaes Bastos. 1996. Drinking water
fluoride levels, dental fluorosis and caries experience in Brazil. Journal of Public Health
Dentistry, Vol. 56 (4): 226-228.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
Dental fluorosis and dental caries
Cross sectional survey
Brazil: Olho D'Agua (Ceara), Vitoria (Espirito Santo) andMaceio (Alagoas)
communities; 457 local schoolchildren aged 6-12 years old (47% males and 53%
females). Participating schools selected for similarities in socioeconomic profiles.
None
The children were required to be life-time residents of the area to be included in the
study.
Children were chosen from these specific three areas because of the known differences
in water fluoride content. Olho D'Agua had water fluoride levels of 2-3 ppm F, Vitoria
0.7 ppm F and Maceio less than 0.01 ppm F. While the areas had similar
socioeconomic status, Olho D'Agua was a more rural community while Maceio and
Vitoria were more urban. Water in Vitoria was artificially fluoridated since 1982 to a
level of 0.7 ppm; water in Olho D'Agua came from bore holes and/or shallow or deep
wells and it was known that the F content varied depending on local rainfall. No
additional information about water in Maceio was reported.
Only exposure to fluoride in the drinking water was considered in the study; no
consumption rate or dose estimates were reported.
Analytical methods on how the fluoride (F) levels were determined were not reported.
The study stated that water was regularly monitored in Vitoria with the majority of the
samples reported between 0.6 and 0.7 ppm F. The level of F in Olho D'Agua supplies
was stated to be between 2-3 ppm but would vary depending on the rainfall levels. The
study did not indicate how often the wells were monitored. No information on how the
levels of F in Maceio were determined or data on any other parameters measured were
provided.
Children were chosen from a low, medium and high fluoride level region in Brazil. The
population of children were chosen from participating schools having similar
socioeconomic backgrounds. Parental consent was obtained for the children chosen for
the study and a questionnaire was included to provide residential history. All
examinations were done by the same person under natural light using plane dental
mirrors. Dental caries was recorded at the level of cavitation. In the permanent
dentition examination, caries was recorded for six teeth (the upper central incisors and
first molars). For the primary dentition examination, all teeth were examined.
Photographs were then taken and used for assessment of the degree of fluorosis of the
upper central incisors. The TF index (Thylstrup and Fejerskov 1978) was used. Color
scale cards were photographed with each film to ensure consistency of film
development. Slides were scored blindly in random order using a graphics light box
without magnification.
For permanent dentition, dental caries was scored with the DMF notation; for primary
dentition, dental caries was scored with the draft notation. The TF index (Thylstrup and
Fejerskov 1978) was used to record any enamel hypomineralization with scores ranging
from 1 to 4 with the higher scores indicating gradually increasing quantitative loss of
49
January, 2008
-------
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
enamel (pitting). The TF scores were determined by photographs that were taken
following the method of Ellwood and O'Mullane (1995).
PROFILER'S NOTE: Although the study states it uses the TF index (Thylstrup and
Fejerskov 1978), a table in the study ranked the fluorosis based on the following TF
scores: 0, 1-2, 34 and > 5 so the profiler is unsure if >5 actually means 5-9, as 9 is the
upper limit of the TF index.
Kappa scores were developed for the reproducibility of recording dental caries and for
tooth TF scores. Comparisons of caries levels were done by using multiple analysis of
variance. No other information on statistics used was reported.
Repeat examinations were performed on 24 subjects to reproduce the recording of
dental caries. Kappa scores were greater than 0.95. Also, re-examination of 25% of the
photographs was performed to ensure the same TF scores were obtained. The Kappa
score for this exercise was 0.85.
Tables 1 and 2 are copied directly from Cortes et al. (1996) and show the mean caries
experience as compared with the TF scores. Caries experience for the six permanent
teeth (DMFT) and for primary (draft) scores presented in the tables were adjusted for
the mean age of the total participants (10 years). In the primary dentition, there was a
statistically significant (pO.OOl) decrease in caries with increasing levels of fluoride. In
the permanent dentition, the incidence of caries in Vitoria was statistically significantly
(pO.Ol) less than the other two regions. In Olho D' Agua, there was a statistically
significant (p<0.05) increase in the mean caries DMFT in those children with TF scores
of 3 or greater when compared to those with TF scores less than 3.
TABLE i
Ms C»=28)
3-1 0.3(0,7} 1.1(1.4}
tn=9)
S5 0 1.3(1.1}
(n=l) (n=!8)
PROFILER'S NOTE: Statistical significance is not clearly shown especially on Table
1. The study report did not provide any reference that the examiner used for deriving the
50
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/date:
DFG/12-06
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
DMFT and draft scores for evaluation of caries. The high limit for the TF score was not
established. The TF index usually goes to 9 while this report only states it as >5.
Cortes et al. (1996) concluded that the overall caries prevalence in these children was
lower than expected but felt that it was because only six teeth were examined in the
permanent dentition, making comparisons easier but underestimating the caries
incidence. The authors also stated that increasing the fluoride levels above 0.7 ppm was
beneficial in reducing caries prevalence for primary dentition but did not appear to be
beneficial to the permanent dentition. In the study, children in the high fluoride areas
with a TF score of 3 or greater had higher levels of dental caries suggesting that if
fluoride intake is too high, severe enamel hypomineralization takes place
increasing the risk of caries.
Ellwood, R.P. and D.M. O'Mullane. 1995. Dental enamel opacities in three groups with
varying levels of fluoride in their drinking water. Caries Res., 29: 137-42.
The study needs more information about statistical analysis performed to allow more
definitive trends to be stated. There does appear to be an increase in caries in Olho
D'Agua when the TF score was greater than or equal to 3. Cortes etal. (1996) does
state, however, that this was a more rural community and some socioeconomic factors
(i.e. limited access to fluoridated toothpaste) may have been additional factors beside
water fluoride levels. The study did not state on how the fluoride levels were measured
in the water and did not provide insight to the reference or method used to score for
draft or DMFT.
Table 3 compares level of fluorosis in 6 permanent teeth (TF index) with caries
(DMFT).
Not suitable for development of a NOAEL for fluorosis or caries prevalence due to
incomplete data.
Not suitable for development of a LOAEL for fluorosis or caries prevalence due to
incomplete data.
Not suitable ( ), Poor (X), Medium ( ), Strong ( )
Nevertheless, observed associations between the TF index score and DMFT are
summarized (Table 3).
Caries experience and TF scores as related to levels of fluoride in water and as related
to each other.
51
January, 2008
-------
Dean, H.T. 1938. Endemic fluorosis and its relation to dental caries. Public Health Rep 53(33)
(Republished in Public Health Rep, 2006 Supplement 1,121: 213-19.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT
ANALYTICAL
METHODS:
STUDY DESIGN
Dental caries and fluorosis
Cross sectional survey
PROFILER'S NOTE: The paper by Dean (1938) is a conglomeration of several studies in
which Dean was either a participant in or principal investigator of and combines data from
these to make conclusions of the relationship between mottled enamel, prevalence of caries
and the amount of fluoride in the water.
U.S.: A 1933-1934 study conducted by the U. S. Public Health Service (PHS) performed a
dental survey of school children ages 6-14 years from 26 states and included a total of 34,283
examinations of white children in South Dakota, 15,465 in Colorado and 48,628 in
Wisconsin.
US/ South Dakota: During April-May 1938, Dean examined 3,300 school children in 51
communities for mottled enamel. All South Dakota counties listed in the earlier PHS study in
which 35% or more of the estimated population of ages 6-14 years had been examined, were
selected for the study.
Dean also used the PHS study data to analyze the percentages of caries-free children in 6
cities but limited this analysis to 9 year olds.
None.
Children in the U.S. Public Health Service survey were 6-14 years old, were born and had
always resided in their respective communities, and had continuous access to a common
water supply used for both drinking and cooking. The PHS Study was conducted from 1933
to 1934 and the Dean study April-May, 1938.
PHS study: 34,283 children ages 6-14 years old from 26 states were included and the fluoride
level in the water was measured.
Dean's 1938 study: All South Dakota counties listed in the PHS study in which 35% or more
of the estimated population of ages 6-14 years had been examined, were selected for the
study. On the basis of the mottled enamel data, these counties were divided into three groups:
a) counties where mottled enamel was prevalent; b) counties where mottled enamel
distribution was uneven; and c) counties which were entirely free from mottled enamel.
Fluoride in water only was measured.
None provided. Consecutive monthly water samples were received from each of the cities
surveyed in the PHS study, which permitted the computation of an arithmetic mean annual
fluoride content of the communal water supply.
PHS study: The study was begun to determine the minimal threshold of toxicity of chronic
endemic dental fluorosis. In some cites, in addition to recording the degree of severity of
mottled enamel, the children were examined by Dean for other enamel defects such as present
caries, past caries, pits and fissures and hypoplasias. These examinations were made in
natural light with the children facing a window using mouth mirrors and new explorers.
Residence and continual uses of the community water were verified by interviewing the
child's parents.
Dean 1938 study: A survey of mottled enamel was conducted in 1938 on school children in
52
January, 2008
-------
51 communities of South Dakota. Mouth mirrors and new explorers were used in all
examinations and children were examined in natural light. The amount of caries was
determined by combining the data associated with the following items: "caries, permanent
teeth", "extraction indicated, permanent teeth", "filled permanent teeth" and "extracted
permanent teeth". For each of these terms, the PHS study provided the number of carious
permanent teeth per 100 children. Adjustment was made for sex, and the amount of caries for
each county was expressed in terms of the number of carious permanent teeth per 100
children.
PARAMETERS
MONITORED:
Dean 1938 study: In addition to definite cavitation, defects in the enamel on caries-
susceptible surfaces showing either a discoloration or an opacity around the edges and in
which the explorer would cling were counted as caries, making the amount of caries appear
higher than usual. To compute an index to show differences in caries in the counties, the
amount of caries (severity) was presented in terms of the number of carious permanent teeth
per 100 children using the 12-14 year old age group.
PROFILER'S NOTE: Because these studies were performed prior to the developed standards
used today for evaluating degrees of fluorosis and caries, there are no exact measurements for
assessment.
STATISTICAL
METHODS:
Statistical analysis was not provided.
RESULTS:
Because the study interchanged data from all of the analyses, the conclusions will be
presented together. The prevalence of caries was determined using a selected group of 9-year-
olds (see Table 2 copied directly from Dean, 1938). The article indicated that observations
showed that a high percentage of children were caries-free in the places with appreciable
amounts of fluoride.
TiWs? 1 ~
« of taiirt tit-e duldi^u M ^eais of &%K in tt »e!«ded ahe-* « h titled itienrdingte their
toiifimwou" n^e yi v* ater ot dilEerent fhmude iFi criueiiti \tnrn
PROFILER'S NOTE: Although the title of Table 2 indicates that it contains information on
the percentages of caries-free teeth, these data are not included in the table, and additional
information from the text must be used to verify Dean's conclusions.
In one part of the text Dean states that of the 114 children exposed continuously to water with
fluoride concentrations of 0.6 to 1.5 ppm, only 5% were caries free, and of the 122 children
exposed to water with fluoride concentrations of 1.7 to 2.5 ppm, 22 (27%) were caries free.
However, in another part of the text, Dean states that for permanent teeth, 60 of the 122
children (49%) were caries free, and the incidence of mottled enamel was 53%. It is unclear
which of these statements is correct.
Evidence of the relation of dental caries to endemic fluorosis was provided by using a
computation of the dental caries attack rate on permanent teeth of 12-14 year olds using the
data from the PHS study correlated with the data obtained on the geographical distribution of
mottled enamel in South Dakota (Dean, 1938). In the group of counties where mottled
enamel was generally prevalent, there were 201 carious permanent teeth per 100 children out
53
January, 2008
-------
of 1,902 12-14 year olds. In the intermediate counties, the examination of 2,765 children
showed 314 permanent teeth affected per 100 children and finally in the counties where no
mottled enamel was found, the examination of 3,481 children indicated 415 carious
permanent teeth per 100 children. Table 3, copied directly from Dean (1938), shows these
results.
Table 3, - Dentil C,«i« Attack rates in peimanent teeth of 12-ii year old white children in
selected South Dakota counties and cities classified according to the prevalence of mottled enamel.
t \ HO ^ (j UU ""' ?£ ir t I'if I, '• ^ M H
• . ,t ii - l!urf
I'-KW t, ' ,* *, er«lwnV
1 y i r __
1 il
to! 'i 'lirlil •'' >l|r I
hf ^ '^ [ i " i \ n, iJ fns^it
, ,
:i«M"i * *.N4Ut.L IHall'.lltl H'lfc 13 I J
Kings bury.
Day........
HughesL,,,.,
MePhersoBu
IM
3t •
Total.
»>'.
sui : si
A!i»'»« fttui Wtfiun>l I, HUP,
M&fttfri, fiiiiit!ic-'l ^risund >fiJ'lcFS0y afni ruf^l fl^l
riurU^'rf' It^rt rif t'fuititj",
i>(.-ltil, Ulna vark-ii Iroiju.-il*, Kaj't-ro'T, LSRH
I.itku J'rc'.t.m »re i>ndnmw. JJe^mof soil &rflu»*l<>F),
!»'•! i;iit''^t I'fiiu'iiiisit es sn WMHily *tr WVH'HV
Si w.e wmftlt'** nr-:ntt6l ia svtremn *"(filj',
li inf'.'tl "MtM'.t Su (.")Tnrii I \'tvTK Jl ^1 > % PR »* CN IlK'fl ! i f l<
<% iL«r hum
N'lVt'Ui- tu miiitSt1'! fT-Hii >•' obM'ri
J L " } IK IT », U a 4i^ ^ vu^tl kt ~ M f/ '^ ^ ^
'Vn t • • ."'1 i if iiMf'U'tt t'imrii 1 m thi- I'lUutp. H»TWld
IH'^'I '<\«' 'n "icru \
Nun 1'ik .if nint 'it'll «>r tnM in (do r.-mitjr.
•I .-' in " r ; •>' ,tf ct.BB-
i ., t . « t - , it. f
PROFILER'S NOTE: While the trend toward less carious permanent teeth with increasing
incidence of mottled enamel was observed with this data, listing the actual fluoride content of
the water in these areas would have been useful.
A similar comparison was performed with data from four cities in Colorado and eight
54
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/Date
VAD/01-09-
Wisconsin cities. These data showed less caries in the areas with the higher fluoride
concentration in the water and more mottled enamel. In Colorado, the non-endemic areas for
mottled enamel, Pueblo, Fort Collins and Denver, the incidence of caries in permanent teeth
per 100 children were 194, 296 and 343, respectively and in the endemic area, it was 163 per
100 children. In Wisconsin, the non-endemic areas for mottled enamel with lower water
fluoride levels incidence of caries in permanent teeth per 100 children ranged from 646 to
917 while and in the endemic area, it was 275 per 100 children. Table 4 is copied directly
from Dean (1938) to provide this information.
Table 4. - Dental Caries attack rates in permanent teeth of '12-14 year old
white children of ALL Colorado and Wisconsin cities listed in public Health Bulletin no, 226
Citj
33ffis**r
fort C^lltM
iiw*» »a
^f B.^a^W'^r
Tvd iU^wff
%f ilwtwkw
Westt Ulil
B»«b«> , ..„„.,..„, ,...,.- ,.„ „,,
I
i \
*SS. P*^™* 3 *> f U11
«*l *2 It .15 ill)
» f W 83 (131
4«5 i!7 , 13 (131
MO mi :> i (Dj
!!9 73J ' 91 U3i
4~ T«h Q 32 fi|,j&
1 "Extraction indicated" for boys "Unknown", 4.2 rate for girls used in this adjustment. - Author.
1 Determination made by Senior Chemist E. Elvolve, Division of Chemistry, Xational Institute of Health.
Approximately the same amount has been reported by DeWitt and Nichols (f . Am. Water Works Assoc.,
29:980-984 (July 1937). Note- For the mineral constituents, other than fluorine, of these Wisconsin waters,
see Public Water Supplies of Wisconsin. Wisconsin State Board of Health, July 1935.
PROFILER'S NOTE: While the data above does indicate the carious incidence as related to
the fluoride content in the water, the data on the incidence of mottled enamel were not
provided.
The study author concluded there was evidence to support the hypothesis that a limited
immunity from dental caries was operative among school children residing in endemic
mottled enamel areas. Also, examinations of 9-year olds showed that a higher percentage of
children were caries free in the communities with a higher water fluoride content.
Dean, H.T. and E. Elvove. 1935. Studies on the minimal threshold of the dental sign of
chronic endemic fluorosis (mottled enamel). Pub Health Rep 50:1719-1729 (Dec. 6,
1935). (Reprint No. 1721).
Dean, H.T. andE. Elvove. 1936. Some epidemiological aspects of chronic endemic dental
fluorosis. Am J Pub Health 26:567-575 (June 1936).
Dean, H.T. and E. Elvove. 1937. Further studied on the minimal threshold of chronic
endemic dental fluorosis. Pub Health Rep 52:1249-1264 (Sept. 10, 1937) (Reprint
No. 1857).
U.S. Public Health Service. 1936. Dental survey of school children, ages 6-14 years, made in
1933-34 in 26 States. Public Health Bulletin No. 226, Washington, D.C., May 1936.
Details on the study conduct, including methods and standards used to assess fluorosis, were
not provided; therefore, evaluation of the results and conclusions is difficult.
55
January, 2008
-------
07
PROFILER'S ESTEM.
NOAEL
PROFILER'S ESTEM.
LOAEL
SUITABILITY FOR DOSE
RESPONSE MODELING
CRITICAL EFFECTS:
The technical reviewer agrees with the conclusions of the profiler. The article provided a lot
of data that were important to set the basis for developing standards in regards to fluorosis
and caries as related to fluoride exposure; however, the reviewer found the information
presented in an unorganized manner making interpretation of the article difficult to follow.
Some of the conclusions made by Dean could not be verified based on the data provided.
The study design did not estimate a NOAEL.
The study design did not estimate a LOAEL.
Not suitable,(X_); Poor Q; Medium Q; Strong ( )
Although not suitable for a dose-response and interpretation of the data were difficult, the
study does serve as a relative source contribution to the inverse relationship between water
fluoride concentration, the incidence of mottled enamel and the incidence of caries.
Dental fluorosis and caries
56
January, 2008
-------
Dean, H.T. 1942. The investigation of physiological effects by the epidemiology method, in
Fluoride and Dental Health, Publ. Amer. Assoc Advanc. Sci., No. 19. pp 23-31.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL
POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL
METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL
METHODS:
RESULTS:
Dental fluorosis
Dental fluorosis
Cross-sectional study of fluorosis and retrospective study of fluoride levels in drinking
water.
US, 5824 Caucasian children in 22 cities in 10 states in several geographic regions; age 9-14
yrs old or in grades 2-12 (see Table 1 below in Results section)
None
Not stated when the individual examinations were conducted.
The exposure groups are described in Table 1 below in the "Results Section)" copied
directly from Dean (1942).)
Drinking water was the only exposure route evaluated. The study author notes, however,
that other factors such as amount of water consumed, dietary and culinary habits, and
probable climatological influences would also influence fluoride intake.
The Elvove (1933) method was used to determine the fluoride concentrations in drinking
water; the sensitivity of the method was reported to be about 0. 1 ppm.
The incidence and severity of dental fluorosis was evaluated in Caucasian children
approximately 9-16 yrs of age in 22 US cites. Fluorosis was evaluated using Dean's Index
of Fluorosis (see Section 2). The fluorosis classification for each child was based on the
degree of fluorosis recorded for two or more teeth. A community Index of Fluorosis was
calculated as the sum of the products of the frequency and fluorosis score divided by the
total size of the study population. The Community Indices of Fluorosis (GIF) were plotted
graphically against fluoride concentration in drinking water.
Dean's fluorosis index was used to evaluate each tooth in the mouth of each study
participant. An Index of fluorosis was calculated for each community; the Index was
computed as the sum of the products of the frequency times each fluorosis score divided by
the total size of the study population.
Community drinking water concentration was calculated as the arithmetic mean of 12
monthly samples.
The size and age (or grade level) of the study populations, the community drinking water
fluoride concentration, the Community Index of Fluorosis, and the incidence and severity of
fluorosis are presented in Table 1, copied directly from Dean (1942).
57
January, 2008
-------
TABU; i
VAIUTION is THE Oners of Pwomw ISWWIOM AUS-OCLM-KU wiTHtiu:<.'«m'wurM Utoiop DCUIMTIC
WATERS of DtfTsagsrt Pwouiw, I'OX<:KXT«.ITHI,V
(ObwrratbuM on S8SI white rfiililmt of ij VLiticN at K» a
Pl:iet
WMfcfcgfi^, IliimwM .
iliehlgiin City,
iHiltawa .,„
SSiincHrilte, Oliie .»,..
Llmrtj Qiao
Marion, Ohio
Elgin. IllinnM
PutWo, Colorado
Kto'tnii'e, tlltBota .„
Aurora, IlHusin
Jalirt, UllMoU „
Elrahttvat, Illitwb ....
QalMburg, XlUnolv ..
tfovii, New M«shco ,-
, Catofmta Spring*,
Colorado
Hriirn'tew, T«XM
iiMttrillo, Testa* „..
' Cftuwflj*, ®otitl» Caro-
lina ._
Lebbook, Taxni
Pout, T»«a
Ctntopiv, Kiui«i8
Antony, Iswa
""3
i
"a
x
1
£
42$
iM
459
4.14
2i3
403
«I4
133
«as
447
170
27S
138
404
91
889
SO
180
38
as
31
Bauxite, Arjiuiiajia .,.: 2ij
2
1 ,
1
$*
1
0.3
0.0
1.3
3.2
ti.l
4.2
a.s
12.S
1S.O
SS.J
40.0
4f.fl
71.0
73.8
STJ
•as
8«.2
97.8
100.0
MOJO
100.0
100.0
*
"1
I
**«
o
£
*w
B
^M
0.61
0.01
0.08
0.0»
0.55
0.22
0.17
.0.31
0.38
9.48
tt.«T
0.69
1.4
1JJ
1.8
2.3
2.1
8.7
34
35
3.3
8.4
So
1 2.
2 5r
1 1
If
0.0
0.1
0.2
O.S
0.4
0.5
0.*
0.9
1.8
1,3
14
1,9
8.8
84
3.9
3,i*
4,0
4.4
s.rt
7.«t
8.61
14. H
Pwrpi'iitii^i" iltKtrili
., gHK-t Mttllt
i
K
n.»
87 .a
8S.4
84.1
37,4
«(!.!!
744
32.8
S3.3
10,3
38.3
28.3
13.0
0.4
4,1
3.1
5.1
1.1
0.0
0.0
0.0
0.0
I
II
1.11
:M>
White
lH«t«|IWKl»«t«
««.
i
>
0,2
(Ml
13.1 I 1.5
13,7
a«.s
,» »*
a.a
as.:i ; ;u
21.S
JS.O
31Jt
34.2
»I.S
S7.1
18.0
TH.8
8.3
fi.fi
€.7
1.1
O.U
fl.ll
o.»
It.*
ti.j
in.rt
13.9
22,3
SU.ft
40.3
23.9
4il
34.0
}S.2
2<».4
li*
O.ti
9.2
0.0
2
I
ll.H
«t,U
a* «
O.ll
«».*
0.7
0.3
j,f,
I,!
a. i
M.S
All fluorMe ^«t«r«lwatia«* vwra made by Senior ClMnnint Eliw Elvow,
:,t«.
tltll'H
(liu«nxtniii*
»li.l |.
1
1
H,ll
Hji
n,rt
ii.ii
n.ii
' ('i.i)
it,»
ii.il
n.ij
tl.»i
1.2
J.I
11.11
s.n
23.7
;«,»
SI. 7,
4ft.(J
Si>,(l
10.8
4T.fi
38..1
ittiug
1
i
Or
•rtiMtl grMle
M.H ' 1S-J4TW.
li-14 "
0.0 ! 12-14 "
U.O 13-14 "
(».U 13-14 «'
o.o is-u "
(>.(> : 12-14 "
!M) 12-14 "
0.0
O.l)
12-14 "
13-14 "
O.U JS-14 "
0.0
rt*
LS
S.I
13.2
11.0
17.9
80.S
3S.5
42.8
33 .»
12-14 "
8-11 "
1-2-14 "
9-19 "
9-12 "
9-U "
9-1 i **
4- 8 gimlet
S-IS "
2-ii "
14-19 fn.
Ou-wlo« oiwW»r»tiim. It. 1«
«( 0.4 or («** »i of BO eon-
rtowi nbuv* 0.0 it kegini to
Utglily tuitjortunt t«t not*
thnt i»n lnd*x of Auontli ia low tit aJwut 0.3 ban been (outtil aawx'iatwl with w«i:wknlrty low (tewtal etwies
vxpititonc* rate. '
The correlation of the CFI with the fluoride concentration in drinking water is shown in
figure below copied directly from Dean (1942).
58
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE
NOT FOUND IN NRC
(2006)
PROFILER'S Initials/date
REMARKS DMO
2/14/07
PROFILER'S ESTEM.
NOEL/NOAEL
PROFILER'S ESTEM.
LOEL/ LOAEL
POTENTIAL
SUITABILITY FOR
DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
IMOfX Of DiWTAi. HUOROStS fOUND ASSO-
CIATED WIIHTME CONTINUED 03£ OF THf
COMMUNAL dWES SyPPiV HAVIMS THE
iNOic*Tfo «,uom0r COHCENTRATIOU.
o — N CM *
VARIATION OF INDEX OF DENTAL FLK08OSIS WITH
THE FLUORIDE CONCeNTRATIOM OF THE COMMUNAL WATER 5UPPLV
{OBSERVATIONS ON 5,824 WHITE CHILDREN OF 22 CITIES Of 10 STATES)
_, -„-„„,,,„,_--..- _„.,,. _ _ ,,,„,„, ,. t i
i t 1 1 1 I 1 'If
, •
M* , """"
*
_ •
*
* •
**
«••*** 11
"f» 1 1 f 1 1 * 1 t || (
5*1 2 3 4 5 6 7 8!i4
FlUORIDE (f) CONTENT OF COMWUNAL WATM SUPPtV IN PARTS PER MILLION (PWA)
F««, &.
Dean (1942) concluded that there was a correlation between the prevalence of endemic
fluorosis in a community and fluoride levels in the public water supply, and that the marked
differences in the severity of fluorosis among the groups studied could be demonstrated by
comparing the CFI. The association between the CFI and fluoride drinking water
concentration followed, with "reasonably precision, a rough S-shaped curve.
Elvove, E
1933. Estimation of fluorides in waters. Pub. Health Kept. 48:1219-1222
Although the study did not take into account other sources of fluoride exposure, the study
occurred prior to the widespread use of fluoride supplements and dentifrices. The study
populations only included Caucasian children, so it was not broadly representative of the
entire U.S. population. The study did include populations from many geographic areas in
the U.S., however, not all regions were included such as the desert Southwest and the cold
temperate regions of New England or the upper Midwest.
Based on the data presented in Table 1, the NOAEL/LOAEL boundary for severe fluorosis
(score of 4 in Dean's fluorosis classification) appears to fall between 1.8 and 2.2 ppm.
Based on the data presented in Table 1, the NOAEL/LOAEL boundary for severe fluorosis
(score of 4 in Dean's fluorosis classification) appears to fall between 1.8 and 2.2 ppm.
Not suitable (_ ), Poor ( J, Medium (_ ), Strong ( X)
An adequate number of populations were surveyed and an adequate number of individuals
were studied within each population to provide sufficient data for statistical analysis. The
CFI followed the same trend as that for dental fluorosis.
Dental fluorosis
59
January, 2008
-------
Den Besten, P.K. 1999. Mechanism and timing of fluoride effects on developing enamel. Journal
of Public Health Dentistry. Vol. 59, No. 4: 247-251.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS MONITORED:
STATISTICAL METHODS:
RESULTS:
Mechanisms proposed for the
formation of fluorosis
Pre-secretory stage
Secretory stage
Maturation stage
Effects of fluoride exposure on developing enamel
Assimilation of data- not true survey or study
Data summary drawn from human, animal and cellular studies.
Not applicable
Not applicable
Not applicable
The report states that most animal studies evaluated fluoride exposure through either
the drinking water or food, but at higher levels than humans normally ingest through
drinking water and diet. Cell culture studies were also described.
Not applicable
The article assimilated data from previous studies and used them to help make
predictions as to the effect of fluoride exposure on the different stages of enamel
development.
The report gave results from several studies that affected each stage of enamel
development but did not use a standard index to describe fluorosis.
Not applicable
Two mechanisms are proposed as being influential to the formation of fluorosis. The
first is a systemic effect of fluoride on calcium homeostasis and second, an effect of
fluoride on cell function either directly through interactions with developing
ameloblasts or indirectly through interactions with the extracellular matrix. The first
mechanism, effect on calcium homeostasis, is only applicable in fluoride exposure
resulting in skeletal fluorosis and was not addressed in the Den Besten (1999) paper.
The second mechanism was examined. The main stages of enamel development are the
pre-secretory stage, the secretory stage and the maturation stage; this report tries to
identify which stage(s) is most affected by fluoride exposure.
The pre-secretory stage is when differentiating ameloblasts acquire their phenotype and
prepare to secrete the organic matrix of enamel. Data identified on both the effects of
fluoride on the proliferation of epithelial cells for enamel (precursors to ameloblasts)
and on cell differentiation found no effect at the levels of fluoride commonly found in
human exposures.
The secretory stage is where ameloblasts secrete amelogenin protein to form a protein
matrix over the full thickness of enamel with mineralization quickly following.
Histological changes have been observed in rat enamel after high levels of fluoride.
Fluoride also appears to cause an inhibitory effect on the uptake of amino acids which
in turn could cause a decrease in the amount of enamel matrix proteins secreted.
The maturation stage starts when a rapid loss of amelogenin protein occurs from the
enamel matrix with mineralization continuing until there is a fully mineralized tissue at
the end. The hypo mineralization characteristic of fluorosis appears to be caused
by a delay in the removal of amelogenin in the early maturation stage. These
effects are observed in animals at doses similar to those of humans. Studies also
determined that the duration of exposure and dose of fluoride prior to the early
60
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date:
REMARKS DFG/1-07
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
maturation stage affected the severity of dental fluorosis.
The formation of dental fluorosis is highly dependent on the dose, duration and timing
of fluoride exposure. The early -maturation stage of enamel development appears to be
the most sensitive. Although the risk of fluorosis is less when exposure only occurs
during the secretory stage, the risk greatly increases when exposure occurs in both the
secretory and maturation stages. The risk of fluorosis also appears to be related to the
total cumulative fluoride exposure in the development stages.
The profiler recommends using pages 250-25 1 (references section) of the study report
to identify references.
This study is not suitable for a dose-response estimate, but does provide insight on the
effect of fluoride exposure in each stage of enamel development. Timing and the
duration of fluoride exposure both influence the degree of dental fluorosis.
Data are not suitable for development of a NOAEL.
Data are not suitable for development of a LOAEL.
Not suitable ( X), Poor ( ), Medium ( ), Strong ( )
Suitable for identifying the critical time of exposure.
Dental fluorosis.
61
January, 2008
-------
Driscoll et al. 1983. Prevalence of dental caries and dental fluorosis in areas with optimal and
above-optimal water fluoride concentrations. J. Amer. Dent. Assoc. 107(l):42-47
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
Dental caries and dental fluorosis
Cross-sectional study of dental fluorosis and caries and fluoride levels in drinking water
US/Illinois: 807 schoolchildren (ages 8-16), approximately 46% males and 54%
females. Nonwhites were <5% of the population in all the communities.
None
From birth until age 8-16 yrs, in April 1980
The exposure groups are described in Table 1 copied directly from Driscoll et al. 1983.
The mean fluoride concentrations for each community were obtained by averaging all
available readings for the years 1964 through 1977.
TfcMt 1 » Study uuninunititts 4i»d (he rrl.ilion at then vwtrt fluoride
i:imawfrat»ht> rcatnwiMinfW ujrfimal Hutiridc l*»v»"l -t.linais
Nani* of M«*tl fluon'lr RflslMW *o ajitim«l
rommumly roncenlnMion Ipprol flmiritt*" \fvn\
K.rwar.**' 1 06 1 -
Monmnuth /.in 2 *
Abiflgdipn i W4 3 x
Cimwoiiii ,* KM
)|niva > "''
Ruihnr t! .* IM 4 >
Table Uiuve 4.tP
PROFILER'S NOTE: While the table indicates that the fluoride concentration for
Ipava falls in the 3x optimal range, subsequent reports in this series of studies
demonstrate that the data for Ipava were included in the 4x group.
Drinking water was the only exposure route evaluated. However, a questionnaire was
developed for parents of 8 children in one area who had moderate or severe fluorosis,
asking questions about consumption of dietary fluoride supplements, consumption of
high fluoride infant formulas, and ingestion of unusual amounts of fluoride dentifrice.
The study authors state that the responses to the questionnaires "did not explain the
clinical findings". The responses were not reported.
Not stated
The prevalence of dental caries and dental fluorosis was assessed in U.S. children, ages
8-16, in 7 communities in Illinois where the water supplies contained natural fluoride at
levels of 1, 2, 3, and 4 times the recommended optimal level of 1 ppm for the
geographic area. Examinations for fluorosis and dental caries were carried out, with one
examiner using Dean's Index of Fluorosis (see Section 2) and the other two examiners
using the Tooth Surface Index of Fluorosis, developed by the National Institute of
Dental Research. (Note that the results using this index were still being analyzed by the
authors and were not included in this paper). The DMF surface index was used to
determine the incidence of dental caries and diagnostic criteria corresponded to those
established in 1968 at the American Dental Association's Conference on Clinical
Testing of Cariostatic Agents. The percentage distribution of children with Dean's
fluorosis scores of 0 to 4 was assessed for each water fluoride level, and Community
Fluorosis Index (CFI) scores were also developed, according to water fluoride levels.
For dental caries, the mean decayed, missing, and filled (DMF) surface scores were
examined based on the water fluoride levels.
62
January, 2008
-------
PARAMETERS MONITORED:
STATISTICAL METHODS:
RESULTS:
Caries
Dental fluorosis
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date
REMARKS SBG 3/28/07
PROFILER'S ESTIM.
NOEL/NOAEL
Dental caries was measured using the DMF surface index and dental fluorosis using
Dean's Index of Fluorosis (see Section 2).
Not stated
See Table 3 for mean DMF scores according to the water fluoride levels
lahlr 1 * M«*an flMF surf* a M"an-> tit t,h)Id;«> .u tjtrttmg
tt< water fluomie k:vd— Hilling
IVaiw fliicrMn V* «tf Mr^n f- lMre«'i (fnlcten DM! S ffii'i ^plimaU'*.,
OfHinta. 3,ffi j n
3 *> oj-linieit Hi i.fC 17 .1
1 • %'inwl |*(i 1.4! '," ^
4 ' jptinia: l"»d ilk' 4"s 7
See Table 2 for the distribution of children according to Dean's Index of Fluorosis and
water fluoride level
T»hle2 • Ptfrc«'i)lag*> distribution of rhiltjnm act nrdutg tr» Dean's fluorosio w,!< '•nor
1«¥«J t.hiltirm «l O.S 1 ? ) *
Oplimat 111. 560 M "> .'4 *R 18 06
1 « optima! 1-U IB? JO.r ^31 1M* »t4 43
3 » uplimil l»I 22 1 26,0 li i ]>4H r« 9J
* « Qptlfflint !<•-, li '» 154 rtf ?"»!' "4 ,',',B
The mean caries scores in all three above-optimal fluoride areas were significantly
lower than in the optimal area. The prevalence of dental fluorosis was characteristically
low in the optimal fluoride area. Substantial increases in fluorosis occurred in the
above-optimal fluoride areas, with the condition being most pronounced in the 4x
optimal area.
None
The study did not take into account other sources of fluoride exposure and no statistics
were done on the data. In addition, the study only included one Midwest state, with
primarily a white population, and so it is not broadly representative of the entire U.S.
population.
There is a clear dose-response for severe fluorosis progressing from a 0.6% at the
optimal fluoride level to 22.6% at the 4x fluoride level. The data may be helpful in
determining the magnitude of the increase in severe fluorosis that has resulted from
increased fluoride exposure through non-drinking water sources.
The mean DMFS scores for the 4x fluoride group was substantially less than that for
the optimal fluoride group (2.02 vs. 3.14); however, statistical significance of the
difference was not reported.
Based on the data presented in Table 2, the NOAEL for severe dental fluorosis (Dean's
Index of 4) is below 1.06 ppm, because 0.6% of the study population exhibited severe
fluorosis at this fluoride level.
63
January, 2008
-------
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
Based on the data presented in Table 2, the LOAEL for severe dental fluorosis
Index of 4) is 1.06 ppm because 0.6% of the study population exhibited severe
at this fluoride level.
(Dean's
fluorosis
Not suitable (_)> P°or (_X Medium ( x), Strong (_)
Statistics were not done on the data.
Dental caries, dental fluorosis
64
January, 2008
-------
Driscoll, W.S., S.B.Heifetz, H.S.Horowitz, A. Kingman, R.J Meyers, and E.R. Zimmerman. 1986. Prevalence
of dental caries and dental fluorosis in areas with negligible, optimal, and above-optimal fluoride
concentrations in drinking water. J. Amer. Dental Assoc. 113: 29-33. [NOTE: Two reports have been published
on this survey prior to the present report (Driscoll et al. 1983; Horowitz et al. 1984)].
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
Dental fluorosis and dental caries
Case control, retrospective
US/Illinois, Kewanee, Monmouth, Abingdon, Elmwood, Ipava, Bushnell, and Table
Grove + US/Iowa, Belle Plaine, Durant, Marengo, and Missouri Valley: 1,123 children
aged 8-16 who were lifelong residents of small rural towns in the same climatic zone
with negligible, optimal, or above-optimal natural fluoride levels in their drinking
water. The towns with optimal or higher fluoride levels (807 children) were in western
Illinois, within 75 miles of each other (Kewanee, Monmouth, Abingdon, Elmwood,
Ipava, Bushnell, and Table Grove). The towns with negligible fluoride levels (3 16
children) were in Iowa (Belle Plaine, Durant, Marengo, and Missouri Valley).
Several comparisons were made after sub-grouping the 1, 123 children based on the
fluoride content of their drinking water (negligible, optimal, or 2x, 3x, 4x optimal), or
on their dental fluorosis score (0-4, per Dean's classification). Thus, the control groups
were (1) children with negligible fluoride in their drinking water, who were compared
to those with optimal and/or above-optimal drinking water fluoride [DMFS/child,
community fluoride index, and distribution of fluorosis scores among children and for
all teeth and tooth surfaces]; (2) children with a fluorosis score of zero*, who were
compared to children with fluorosis scores of 0.5-4 [DMFS/child and percent sound,
filled, and decayed teeth]; and (3) children exposed to optimal drinking water fluoride,
who were compared to children exposed to above-optimal drinking water fluoride
(TSIF scores for permanent first molars).
included only children exposed to above-optimal drinking water fluoride levels
Lifetime, which was 8-16 years
Schoolchildren aged 8-16 (grades 3-10); mean age =1 1.48, -46% male and 54%
female, who lived their entire life in one of the 1 1 small rural communities in western
Illinois or in Iowa, and whose primary drinking water source was the public water
system. The current paper (Driscoll et al. 1986) did not identify the 7 Illinois towns in
which 807 of the children lived, and only provided their drinking water concentrations
as optimal, or 2x, 3x, 4x optimal. The identity of the towns [and their mean water
fluoride concentration] were obtained from a previous publication by the same authors
(Driscoll et al. 1983), as follows: Kewanee [1.06 ppm], Monmouth [2.08 ppm],
Abingdon [2.84 ppm], Elmwood [2.89 ppm, Ipava [3.77 ppm], Bushnell [3.84 ppm],
and Table Grove [4.07 ppm]. Based on the geographic location, the study authors
considered 1 ppm to be the optimal water fluoride concentration, and thus classified
water supply for the towns studied as optimal (~1 ppm), 2x optimal (~2 ppm), 3x
optimal (~3 ppm), and 4x optimal (~4 ppm).
The 3 16 children in the four Iowa towns (Belle Plaine, Durant, Marengo, and Missouri
Valley) were also 8-16 years old, with a similar age distribution as the Illinois children
(no further details provided). The children's drinking water fluoride concentrations
were classified as "negligible," and were only quantified as being <0.3 ppm.
No information was provided on other possible sources of fluoride exposure (e.g.
toothpaste, mouth rinses, etc.)
The study did not indicate the method used for measuring the water fluoride
concentrations, or report data on any other water quality parameters.
65
January, 2008
-------
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental caries (DMFS)
Effect of dental fluorosis on
dental caries
In April 1980, teeth were examined from 807 children aged 8-16 who lived in 7 small
rural towns within 75 miles of each other in western Illinois. These children had lived
their whole lives in their respective towns, per a questionnaire completed by their
parents. Informed consent was obtained. The towns had varying levels of fluoride
naturally in their drinking water, which came from deep wells. Based on the
geographic location, the study authors considered 1 ppm to be the optimal water
fluoride concentration, and thus classified water supply for the towns studied as optimal
(~1 ppm), 2x optimal (~2 ppm), 3x optimal (~3 ppm), and 4x optimal (~4 ppm).
Teeth were examined in April of 1980 at the children's schools using portable dental
chairs, artificial lights, plane dental mirrors, and #23 explorers, but radiographs were
not made. Fluorosis was assessed in all children by one dentist using Dean's
classification system, and by two other dentists using TSIF and the criteria of Russell
(1962) to distinguish between fluorosis and nonfluoride enamel opacities. Another
dentist examined each child for dental caries using the DMFS index.
Two years later (April 1982), 3 16 children aged 8-16, of comparable age and gender
distribution, were similarly examined in four small rural towns in Iowa. The towns
were in the same climatic zone as the Illinois towns, but had negligible (i.e., <0.3 ppm)
water fluoride concentrations (towns unavailable in Illinois).
Dental fluorosis was assessed in all children by one dentist using Dean's classification
system, and by two other dentists using TSIF (tooth surface index of fluorosis) and the
criteria of Russell (1962) to distinguish between fluorosis and nonfluoride enamel
opacities. Another dentist examined each child for dental caries using the DMFS
(decayed, missing, and filled permanent surfaces) index. The Community Fluorosis
Index (CFI) was determined by the method of Dean (1942).
Scheffe's method for multiple comparisons was used to compare the mean DMFS
scores of children according to water fluoride levels (negligible and Ix, 2x, 3x, or 4x
optimal), and according to fluorosis classification (0, 0.5, 1-3, and 4). The Community
Fluorosis Index (CFI) of areas with negligible or with optimal water fluoride levels
were compared by the t-test.
The mean DMFS score per child was inversely correlated with their drinking water
fluoride concentration, as shown in Table 1. DMFS scores were statistically
significantly higher in children with negligible drinking water fluoride than in children
with optimal water fluoride, and both of these groups had higher mean DMFS scores
than children from areas with 2x, 3x, or 4x optimal drinking water fluoride. DMFS
scores of the 2x, 3x, and 4x exposure groups were not significantly different from one
another.
fiftlf .'1 •»' Mta«\DMrejea*£»;S§w^f«f to wster-flyprtdejlevtl; \ •'•'•••;.;• 1'- •
' ,' . ';*'•' .. ;'"'•"" , ,'<•••'''' '';''• "'"-'•',' ;,'„ -. - >; - ",'•• '.,"•" • - , .V. "-; ' "iDiffete'tiee''1',- •
Water-Htiojridte1 ;' ,>''>'"'' ,'V';'. * ., 'KiX-pf V - • v^an^MSlIS/,, V.V :s':ftonj';i)feglj|Me
'level"' '• . , children' ;'" " '' pet'cWW . •1-':",>1-- '';(%)'','>'''•
Negligible .! ' . ' 816 . '- . 5,0? '• ". , ; - '''•..'.
Optimal,, , • ,- - • SS6 ••'. ' S,W* • ' :/ , ; '-/• ' 'S8il
Two-rimes optimal,-:. , 143 ' ' • ' t.ffft \ '•' ' .'" . . 61,1 ;
Three-times optimal . ' -'W...^ . ' 1.4If '•''-. ,"'.;.. „ • .72.4
Four-limes optimal, :•', .' -,.,.', J.36-V • .• ''.,?;6?,t- -,'''.- 'V '•".-'."•.' 60,2' '.
*%mfieaMly'leweritan negligible (P * ,OC(!> • , ', '- ' ; '. • • ' '*' ';~ '• ' •. • V'':;;'"
tSffnilteamlj'' lower than- negligible and optimal' (?•< ,01), ' •'• '•'•',' :', ,"''*' ; ,'
A comparison of the mean DMFS scores with the Dean's fluorosis classification of
children from areas with above-optimal water fluoride showed that children with a
fluorosis classification of 4 had statistically significantly greater DMFS scores/child
66
January, 2008
-------
Dental fluorosis
than children with fluorosis classifications of 0.5-3, but the DMFS scores did not differ
significantly among the latter groups (Table 2). Although children living in areas with
negligible or optimal water fluoride concentrations were not included in this
comparison, the DMFS score/child for areas with negligible water fluoride (5.07) was
greater than the DMFS score/child of children with category 4 fluorosis (2.96),
suggesting that even 4x optimal water fluoride protects against dental caries.
Table 2. Mean DMFS scores for children according to fluorosis classifications*
Classification of fluorosis | Number of children | Mean DMFS per child
0 87 1.89
0.5 112 1.40
Ito3 218 1.58
4 54 2.96*
* Data limited to children residing in above-optimal fluoride areas.
* Significantly higher than scores for fluorosis classifications of 0.5 and 1 to 3 (PO.05). All
other differences are not significant.
A comparison of tooth fluorosis scores (Dean's index) with the percentage of teeth that
were sound, decayed, or filled in children from areas with above-optimal water fluoride
concentrations showed that children with a fluorosis classification of 4 had markedly
(2.5 to 7.3-fold) greater percent decayed or filled teeth than children with fluorosis
classifications of 0.5-3, but there was little difference among the latter groups (Table 3).
TMkl|,3 i/p'»re«il*gejtf %i»Mi^f N:«-*»"^t «*»pay«^i ••;0'•,'l''•v^^'^ >.•• „ ''.'• '*'..,* •]./ '} ~ ''•• ^: "•"?", '.,.,."
•FKiorpfe soSre ''"''.• ,"* *';;*-" ' . J . "''Vj '•""'', •', ' • • ' < '• -"I" • ' .,'• .Decayed m •'
,'oftiwflt*--.'" • - - ,"•'••'•' Sound (%J' ,'" ' TpeeayW i%) •• Fpl«d($&) - • 'N fiJ!«M%)|, -. - '
-o "•• -, . ' '•• 'Vsfc. «""-"' ]' ;t't , • ',- 5=8 'r\ '} -..'.-••?,9-~
,5 ...'..', ,; ',.«7.J-.' ' ; •'-•"'-,' 'iVi ''v I& ' ' • ' S.7 ' '
ItaS, , r ,, . • '954 j1"-., '-/ f,8 , , - • ,3.8. -/:- - ,4,0 ..
4 ., '. . . .. ,. ,'•;/.- 'SIM ' •• {, "-.A4.,,".. . ., . • '!'?$'".. •'•', ' .18.6.;
*0@^'sr« limited io ehilElrffl'tiesJdiii^l^'afe^^ckapliM^Vfi^11^^^8' - ""'?'','.-, ! E- '
f Percentages may 'not mmto lf)0'b««iW!'tfr' .MstEiButton offhiwren (S3 •<•.*' ''.'_' ' .Community,
Water»fluo!r|4t'p ' '' •' -v, - -v'-\- ,', ,' +• '•'("'," ^'- ,'-i, iv •',,- u---* A,,', „-'„,"•'-• ',-— '' ' . -' fj normals '
level"' ..-",, , ' •.'-,> .;0< ' ' .-,' fl.5 ;V-, '• 1J- - ,' '.,j "''"•"- ~\8 .'•'"-"" '" '$ ,. •. ' Mex'stor* L" -
Negigible • ;," ' -9j;iq ' '' •'' 4;f' -v ' '1.9 ';: '•),{» ,\ '-<(•:,"'" ''& ' "' ^Oft,'' ,.'
Optimal*1" :'," r '' -W.D.,," 2B.S, ' ,• t.4 / . ' *,&• 1 ' . 1.8,' ^ ;"0;8 ' ' :,3t '••'
*te«ntsif!»'jtayiTOt"«jnvtqJ£H!) because cjf rojjjn'iirig;', ;' _' ""'',' V • "' " ' , '/' :
A similar pattern was seen for the distribution of fluorosis scores on a per tooth basis
(Dean's index) or on a per tooth surface basis (TSIF). Fluorosis was somewhat more
prevalent in teeth from children living in areas of optimal than in areas of negligible
water fluoride concentrations, as shown in Table 5.
67
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
1 Tibia 6 • P»re»nt«a« tttitftultoti ol fluonMli'teom.'for all teeth (DMH'B index*)
•fld tor all te«th aurf«c»t (T$lf) In eo«MiiunttlM'wIth'n«gligib)«'«nd opimaf levels ol
w«tor4luarid«conemtnitlons..-.\ . ':,•'••
J • • ' D«iuif» lade* .. '
" '• , • i '" " . Pfelributjon'pf fluofotii scow* (%)
Water-flutnid* Storof .>•.-••• -•-.•.•• ,....-•_
' fcvd ' • • teeth ' • • . o: •- ' ;8 " . ' i 5 3 • 4
Negligible" ~'6,H9.: • .86,5 ' 8.8 ' - I,J ' 2 fl Q
.Qpilnwl • . - 6,'288 :, ,-/..79;7 ' '"'lS,f;" 4'g - 1.5 ,8 ,I> ' -
',;.•• , '•••"..'•• rsaf. • , ,.' '••
1 ' . ' ' , • , Bbtribiitten of fluorosis scores (%)
-'W*cer*fluoridt No. «('• '; i1 ' ,' :
' • 'evel tooth surfaces 0 l ,2 345 6 • ' 7
Negligible • 14.78S .. 94.1 -6.4 ' JS .0 o' 0 0"' 0-
• Optimal 16,'SW 84.6 it,4 . 2.0 U 0 0* • - 0 0
' , *Four affecwd surfHepR,
A comparison of the distribution of TSIF scores of permanent first molars of children
aged 8-10 and 13-16 was made according to type of tooth surface (occlusal, buccal, or
lingual) and water fluoride level (optimal vs. above-optimal) (Table 6). TSIF scores
were greater in molars exposed to above-optimal fluoride levels than molars exposed to
optimal fluoride levels for all tooth surface types, for both age groups. However, TSIF
scores were lower for the 13-16 year olds than for the 8-10 year olds at both water
fluoride levels, suggesting to the study authors that the older children's molars were less
affected by fluorosis.
I»W« B • Percentage cHatrlbuiion of TSIF «cow» tor f»»r«an»nt flnrt molars In
;bofYim«5ltte» with optimal and abotftt-opllwal water fluoride ItvsH according to type
>f'|o6tK aurtact and age .proiip. .- • • - ' .
',;."'''';;' . , • '""•-; • '- Obtritathm of TStP MOTM (%) ' - '; •-•
' V-fliiorld*""'1' -.:' •• • '•'• ^.'o ''O-y'W-oIdi ; IS- to 16-year-oltb :..,.'
-J?5?— ^— _— __ '^^—^^SJ1^^^^J^^^^L_ ' 1»S • -.4 to' 7'
' • ' _'..'•'' ' ' O«cl»Ml . •
.Ppt'tirti ~i].S'' ' 28,2 - 0 89-6 -' 10.4 • ;0
Abort optimal • iB.» '•-/•' '64.'7 . - '0 " "64.8 3s:Z -'0
•••,.. i - Buccal ' -' . , .
Optima! . .. 'SlS,.-'-\ '•• -1&.8.S- -•-•• 0- ,u''.'/ ,9i;7 - '8;5 • 0
Abow optimal S7.3 '• -,/ 6S.7 .•'•.' .. 0,1' • ,-6'M • ' 38,4 9.2
':' . ' ' • •' .:,.-.'"'. : •'."• ';'•'•, ' •'. , • lLln(pi«l'p' ,
op*510**' . • , ;7i».a..-. ,/--\- -?o;5' ;• • ,:. o1 ••,•' .'«4'.i , 5,9 o
Alxwe optimal *l.6. •''• ,'58,4',-' ,''s,6 - ' ;69.4 88,4 23
- *One lifatied turfiuc, • '.""''., . • ' '
Children from above-optimal fluoride areas had fewer dental caries (i.e., mean
DMFS/child) than those from optimal fluoride areas, who in turn had fewer dental
caries than children from negligible fluoride areas. Children from above-optimal
fluoride areas with severe fluorosis (Dean's index 4) had an increased the prevalence of
dental caries relative to mild and moderate fluorosis (Dean's index 0.5-3); the study
authors speculate that this finding may be due to food, debris, or plaque becoming
trapped in the defective enamel.
The study authors speculated that the reason for the higher fluorosis scores of the
molars of the 8-10 year old group, relative to the 13-16 year old group, could have been
due to higher ingestion of fluoride from dentifrices by the younger group, or to greater
enamel abrasion from more years of tooth brushing, and subsequent remineralisation of
mild fluorosis, by the older group.
The authors countered the findings of Leverett (1982), who proposed that dental
fluorosis has become overly prevalent in the U.S. due to ingestion of fluoride from
sources other than drinking water, particularly in areas with optimal water fluoridation.
68
January, 2008
-------
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/date:
SM 1/8/07
DMO/3/29/07
PROFILER'S ESTIM.
NOEL/NOAEL for fluorosis
and for fluorosis -induced caries
PROFILER'S ESTIM. LOEL/
LOAEL for fluorosis
In contrast to Leverett's finding that 28% of children from fluoridated communities had
mild fluorosis, Driscoll et al. (1986) found that only 14.6% of children in optimal
fluoride areas had definite signs of fluorosis, which was only slightly greater than the
10% estimated by Dean in 1936.
Driscoll et al. 1983. Prevalence of dental caries and dental fluorosis in areas with
optimal and above-optimal water fluoride concentrations. J. Amer. Dental Assoc. 107:
42-47.
Leverett, D.H. 1982. Fluorides and the changing prevalences of dental caries. Science
217: 26-30.
Russell, A.L. 1962. The differential diagnosis of fluoride and nonfluoride enamel
opacities. Public Health Dent. 21: 143-146.
This was a well conducted study showing that fluoridated water protects against dental
caries, but also increases the prevalence of dental fluorosis. Mean DMFS scores per
child were found to be significantly lower at higher fluoride drinking water levels;
however, at the same time children residing in above optimal fluoride areas and who
also had severe fluorosis had a significantly higher mean DMFS score than those with
lower fluorosis scores (NOTE: the mean DMFS difference between children with a
Dean score of 0 and that of 4 was not statistically significant (p=16), but it was still
substantial, i.e., 1.89 vs. 2.96). NOTE: the mean DMFS score for the severe fluorosis
group (2.96) was still below the mean score for the optimal fluoride area (3.14) and
substantially lower than the score for the negligible fluoride area (5.07). These
differences were not analyzed statistically.
Information on fluorosis was provided by comparing the percentage distribution of
children (Dean's index) or teeth (Dean's index and TSIF) for the negligible and optimal
fluoride areas. The results indicate that even in the optimal fluoride areas (1 ppm)
severe fluorosis occurred in a small percentage of children and teeth. The same
information was not provided for the above optimal areas; therefore an overall
evaluation of the dose response relationship for severe fluorosis is not possible.
The study revealed greater fluorosis in first molars of 8-10 year old children than 13-16
year old children, but the reasons for this, as stated by the study authors, were not
conclusively determined.
A NOAEL for minimal fluorosis could not be determined from the study data because
fluorosis occurred even in the negligible fluoride area (in about 7% of the study
population). The NOAEL for severe fluorosis appears to fall between the negligible
fluoride concentration (<0.3 ppm) and the optimal concentration (1 ppm), as no
individuals in the former group had a Dean index of 3 or 4.
Based on the data collected, fluoride appears to have an anti-carries properties even at
levels exceeding optimal. However, when the children from the negligible and optimal
fluoride areas were excluded from the analysis, the lowest DF tooth incidence (was
found among the children with a 0.5 score for fluorosis (2.7%) not the children with no
fluorosis. The highest (19.6%) was found among the children with a fluorosis score of
4. Children having scores of 1-3 had an intermediate incidence of DF teeth (4.5%).
All of the children included in this compilation of the data received water from a source
that was classified as 2x optimal or greater. The mean DMFS value for the children
with a fluorosis score of 4 was significantly higher than those with fluorosis scores of
0.5 or 1-3 but not those with no fluorosis.
A LOAEL was not identified for a minimal level of fluorosis. The LOAEL for severe
fluorosis appears to correspond to the optimal fluoride level of 1 ppm because a small
percentage of children in this group had severe fluorosis (data not analyzed
statistically).
69
January, 2008
-------
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
Not suitable (_ ), Poor Q, Medium (X), Strong (J
The percentage distribution of children (or teeth) with fluorosis was not documented for
all drinking water fluoride levels; therefore, an overall dose-response relationship can
not be established (although this information might be available in earlier or subsequent
studies in this series, i.e., Driscoll et al., 1983). Nevertheless, the distributional data
given for the negligible and optimal fluoride areas might be used in conjunction with
that from earlier studies to assess the relative increase in fluorosis due to the increase in
fluoride exposure from non-drinking water sources.
Dental fluorosis
70
January, 2008
-------
Ekanayake, L. and W. van der Hoek. 2002. Dental caries and developmental defects of enamel in relation to
fluoride levels in drinking water in an arid area of Sri Lanka. Caries Res., 36: 398-404.
Ekanayake, L. and W. van der Hoek. 2003. Prevalence and distribution of enamel defects and dental caries
in a region with different concentrations of fluoride in drinking water in Sri Lanka. International Dental
Journal. 53: 243-248.
[PROFILER'S NOTE: Two sources for the same study were identified; data from both sources were used to
develop the assessment and the tables provided are referenced].
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
Dental caries and defects of enamel
Cross-sectional survey
Asia/Sri Lanka/Uda Walawe: 486 (236 boys and 250 girls) schoolchildren aged 14 years
old from Uda Walawe, a rural area in Sri Lanka, were used in the study.
PROFILER'S NOTE: The exposure groups in Sri Lanka may not be representative of
similar aged groups in the Unites States due to the arid climate of the area (average
rainfall less than 2000 mm per year).
None
1987-2001; Only children that were life-time (14 years) residents of the area were used in
the study.
Water samples from the children's homes were obtained and sampled for fluoride. Most
of the drinking water to the individual homes came from wells. The children were then
divided into 4 groups: <0.3 mg/L; 0.3 1-0.49 mg/L; 0.5-0.7 mg/L or > 0.7 mg/L. In Uda
Walawe, the fluoride content of the water was reported to be 0.05 to 6. 10 mg/L
(unpublished data from the International Water Management Institute in Colombo, Sri
Lanka).
PROFILER' S NOTE: The profiler notes that the highest group of fluoride in the study
was presented as > 0.7 mg/L with no limit to the upper boundary making interpretation
more difficult as to the actual fluoride exposure.
Only exposure to the fluoride in drinking water was assessed in this study. The children
were questioned about tooth brushing frequencies and the type of materials used when
brushing teeth.
Fluoride samples of the water were determined by an Orion ion-specific electrode within
4-5 weeks of obtaining the sample. TISAB buffer controlled the ionic strength of the
water samples and the instrument was standardized using 1 and 10 mg/L fluoride
standards. No other water parameters were measured.
Names of children in Uda Walawe that were 14 years old were obtained from six
government schools. All children within the age group present in school on the day of the
examination were included in the study. Examinations of the children were performed by
one examiner during January and February 2001 and took place outside under natural
light at the children's school. After any debris was removed using a piece of cotton wool,
the teeth were examined for caries and defects in the enamel. Other information gathered
at the examination was: name of child, date of birth, gender, name of school and home
address. Intra-examiner reliability was assessed by re-examining 10% of the students.
71
January, 2008
-------
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental effects of enamel
In April and July 2001, research assistants went to the homes of the children that had
participated and an interview was performed with the mother, father and/or guardian. The
following information was obtained: materials used to brush teeth, age when first started
brushing teeth, source of drinking water and duration of residence. A sample of drinking
water was collected from each home for measurement of fluoride concentration.
PROFILER'S NOTE: While 5 18 children were examined at the schools, in the follow-up
visits, drinking water samples were only collected from 486 homes; therefore, the data
presented are only for the 486 children.
Caries were determined by using the methods of WHO (1997). Developmental defects of
enamel were determined using the modified DDE index (Clarkson and O'Mullane 1989)
onbuccal and labial surfaces of 10 teeth.
PROFILER'S NOTE: The NRC (2006) stated that the DDE index by Clarkson and
O'Mullane emphasizes only aesthetic concerns and is not technically an index of enamel
fluorosis.
Data were analyzed using the Stata Release 6 (1997) software package. Non-parametric
tests were used: Mann- Whitney U test to compare mean DMFS values in 2 samples and
Kruskal-Wallis test to compare mean DMFS values in more than 2 samples. The
differences between proportions were determined by chi-square test and Fisher's exact
two-tail test. The association between the different categories of diffuse opacities in the
DDE index and DMFS values was determined by Spearman's rank correlation. Cohen's
kappa statistic was used to determine the intra-examiner reliability.
Table 1 and 2 are copied directly from Ekanayake and van der Hoek (2003) and show the
enamel defect and diffuse opacities when compared to the level of fluoride in the water,
and also the mean number of teeth affected. Table 1 shows a statistically significant
(pO.OO 1) increase in the percentage of children with enamel defects and diffuse opacities
with the increase in the fluoride concentration of the water. Fifty seven percent of the
children exposed to fluoride levels of >0.70 mg/L had enamel defects as compared to only
29% of those exposed to < 0.30 mg/L. Table 2 also shows a statistically significant (p< 0.
0001) increase in the number of teeth affected with enamel defects in each child as the
fluoride levels went up. The study also identified the maxillary first premolar as the tooth
most often affected. Data for the study were combined for both genders as no differences
were identified. Also, over 75% of the children had used fluoridated toothpaste from the
ages of 9-12 months on.
72
January, 2008
-------
Dental caries
Table 1 Percentage distribuiion of subjects with different types of enamel defects
according to fluoride group
Fluoride group
Type of defect <0.3 mg/l 0.31-0.49 mg/i 0,5-0.7 mg/I >0.7 mg/I
(n=119) (n=126) (n=88) {n=153}
Any defect" 29,0 35.0 43,0 • 57.0
Demarcated opacities 3,0 2.0 3,0 3,0
demarcated white 3,0 0.8 2.0 1,0
demarcated yellow 2.0 0.8 1.0 0.0
Diffuse opacities" 27.0 35,0 41.0 55.0
diffuse line 4,0 6.0 2,0 3,0'
diffuse patchy 12.0 19,0 20.0 16,0
diffuse confluent"" 19,0 - 22,0 27.0 46.0
diffuse mixed"" 9,0 10.0 14.0 22,0
Hypoplasia 0.8 0,0 0.0 0.0
•£*=24.26, peO.001; ' "xs=23.97, p<0.001 ; " X2=28,55, p«O.OQ1 ;
""¥=10,8,0=0,013
A subject may appear in more than one category of enamel defects
Table 2 Mean number of teeth with different types of enamel defects per subject
according to fluoride group
Fluoride group
Type of defect <0.3mg/l Q.31~0,4img/l O.S-0,?mg/l >0.7rng/l
(n=11S) fn=126) (n=88) (n=183)
Any defect' 1.60±3.0 1.91*3.2 2.22*3,2 3.61 ±3.9
Diffuse opacities' 1.56±3.0 1.89±3.2 2.1?±3,2 3.58*3.9
Demarcated opacities 0.03*0,2 0.02±0.1 O.OStO.3 0,03*0.2
Hypoplasia 0,01*0,01 0,0 0.0 0.0
difference between groups significant at p<0.00Q1
PROFILER'S NOTE: The study authors (Ekanayake and van der Hoek) used the DDE
index by Clarkson and O'Mullane (1989) and the NRC (2006) does not endorse this
method as a true indicator of fluorosis, although the number of defects and diffuse
opacities did appear to increase with higher fluoride concentrations. The upper boundary
to the fluoride exposure is not known as it is only presented at >0.7 mg/L.
The next two tables are provided and copied directly from Ekanayake and van der Hoek
(2002) showing the caries prevalence and mean DMFS in regards to the number of diffuse
opacities and DDE index score. Table 2 below shows the caries prevalence and the mean
DMFS in children with and without diffuse opacities and indicates there was a higher
incidence of caries in children with opacities but only statistically higher in the >0.7 mg/L
group. In Table 3, pooled data indicate a significant increase in caries prevalence in those
children as with the highest DDE score (6) and mean DMFS (1.04).
73
January, 2008
-------
Table 2. Canes prevalence and mean
DMFS in children with and without diffuse
opacities h> level of fluoride in tlnnksn^
wafer
Fluoride group
Diffuse opacities
Caries prevalence
Mean DMFS
O.SO
-4X?Oms/t
Total sample
,*«nt*ii = ST|
present |n = }2 1
plOM-'nt (11 - 441
mln »52!
tut(n - 3ft)
l«Virnttn X4J
abxwlln- 290)
pit-sent (n = 1%)
t>
IX
18
2?
9 25
14 20
5(1 36
p - 0 UJh
60 21
57 29
p = 0.034
OJ7± .2
0.65 ±1.8
052tt.4
O.«j9±1.5
0,42
0.51 ±1.5
0.73±1.5
p = O-'C''
The number of subjects with and without diffuse opacities in the. different fluoride groups
is giwn in parentheses,
Tatolo 3, .Association between the severity
of developmental defects of enamel, caries
prevalence and mean DMFS in the
different fluoride groups
Fluoride group
ODE indei seote
Cute pitv
,; {4MB 3)MMjj
iH3(hng/i
•»0"Umi/l
Total sample
The number of s
theses.
DDfc -Mite (I In = h 1}
DDE«are3t4tn-7)
Dr>FsaircMn= 14)
DDF score hfo* 111
DOl'suire X«"tln- I4
DF* score 5(0*171
l)l>E«.-»n-*ln- HI
DDF. win- 1+4 In- III
DDi noieJln- 1J1
DDi'saacidj- 12t
DDFscorcOlnssh'J)
DDEs, ore 3+4 (n = K>
nUFv-wOln* Iwni
DDEsccre3+4m«40)
0PP:*M.ri:5lii*1?i
DDE »wi- 6 IB = 691
16
0
14
1
18
U
14
II
21*
31
18
15
42
20
13
37
39
21
IS
29
38
fljhtl.:
OCK)
029±01
1 5ti±2.R
0 7 1 1 1 9
OWsl.4
0.64 s If
046il.;
HMIil.g
042±)0
fl,25±O.T
OS6*I.5
1,06*1.1
051,* 15
0.3l)*0.>»
Hrt.t 1 S
mg u> the higbfst UI>I MMte ncordcd is given in paren-
Profiler's Note: If the concentration data are removed and one just looks at the incidence
of cavities versus DDE score, the relationship is significant (p= 0.009) and shows the u-
shape curve one would expect. The u-shape becomes even more pronounced if 5 and 6
scores are combined as were the scores of 3-4.
The next Table 3 is from Ekanyake and van der Hoek (2003) and shows caries prevalence
and mean caries experience according to the fluoride groups but does not indicate any
statistical significance between groups.
74
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER' DFG/12-06
S
REMARKS
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
Tabte 3 Caries prevalence and mean caries experience according to fluoride group
Fluoride group
<0,3mg/I 0.31-0.49mg/I 0.5-0. 7mg/l >0.7mg/l
(n=119) (n=l28) ' (n=88) (n=153)
Caries prevalence 10% 24% 24% 25%
MeanDMFT 0,29 ±0,7 0.35 ± 0.8 0.35 ± 0.7 0.54 ±1.C
MeanDMFS 0.45*1.4 0.60*1.7 0.67 ±1.7 0.67*1.3
MeanDS 0.21 ± 0.5 0.27 ±0.6 0.27 ± 0.6 0.44 ± 1.C
Mean MS 0,17 ±0.4 0.28 ±1.3 0.34*1.5 0.13 ± 0.8
MeanFS 0.07 ±1.4 0.05*0.3 0.06 ± 0.3 0.10 ±0.4
DMFS (occiusal) 0.23 * 0.7 0.25 ± 0.8 0.27 ± 0.7 0.33 ± 0.9
DM FS (mesial/distal) 0.07 ± 0.5 0.11*0.5 0.15*0.6 0.07 ± 0.3
DMFS (byccal/palatai) 0.15 ±0,6 0.24 ±0,7 0.26 ± 0.7 0.28*0.8
PROFILER'S NOTE: Data provided from the earlier paper (Ekanayake and van der
Hoek 2002) that used the DDE index indicated significance in caries prevalence when
compared to the amount of fluorosis as seen in the first Table 3 listed, but the second chart
(Table 3 in the 2003 article) indicates there was no significant increase in caries
prevalence when it was evaluated by water fluoride group and mean DMFT and DMFS
score.
In the first paper (2002), Ekanayake and van der Hoek concluded that the relationship
between fluoride levels in drinking water, diffuse opacities and caries suggested that the
appropriate level of fluoride in drinking water in arid regions of Sri Lanka should be
around 0.3 mg/L. Also, individuals with severe forms of enamel defects in high-flouride
levels were susceptible to dental caries.
In the second paper (2003), Ekanayake and van der Hoek revealed that 29-57% of the
children were affected by developmental defects of enamel and the caries prevalence
ranged from 18-25% in the different fluoride groups indicating wide differences in the
prevalence and severity of enamel defects and wide variations in the individual responses
to high fluoride levels in the water. They also stated that the study demonstrated a need to
look at other factors that contribute to enamel defects. .
There were some deficiencies in this study. The study (Ekanayake and van der Hoek 2002
and 2003) used the DDE index according to Clarkson and O'Mullane (1989) for
determination of fluorosis which the NRC (2006) felt was not an appropriate index to be
used. The study did not account for any other types of fluoride besides the usage of
fluoridated toothpaste that could have added to the fluoride exposure. Because of the
higher temperatures and arid conditions of this area, the profiler questions the
applicability of this data for use in the United States. The study also did not give a limit to
the highest exposed group and the profiler does not know what percentage of children
could have been exposed to very high levels.
Data are not suitable to determine a NOAEL for fluorosis or caries prevalence.
Data are not suitable to determine a LOAEL for fluorosis or caries prevalence.
Not suitable (X), Poor ( ), Medium ( ), Strong ( )
However, the study contributes to the weight-of-evidence for the relationship between
75
January, 2008
-------
MODELING:
CRITICAL EFFECT(S):
fluorosis and cavities.
Fluorosis and caries prevalence
76
January, 2008
-------
Eklund, S.A., A.L Isamil, B.A. Burt, and J.J. Calderone. 1987. High-fluoride drinking water,
fluorosis, and dental caries in adults. J. Amer. Dental Assoc. 114: 324-328.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
Dental fluorosis and cornonal caries characterized by DMFT (decayed, missing, and
filled permanent teeth index)
Case control, retrospective
USA/New Mexico/Lordsburg: Lifelong adult residents (n=164) from Lordsburg, NM
(60 miles from Deming, NM), which has naturally fluoridated drinking water
containing 3.5 mg/L fluoride. The adults were whites of Hispanic or other origin
(89.6% Hispanic); 43.2 years old; 67.1% female; had 11.4 years of education, 27.0
teeth (mean values). Income level was similar to the control population (N = 187
examined, which generated 164 subjects). Same water supply has been in use since
early 1900s.
USA/New Mexico/Deming: Lifelong adult residents (n=151) fromDeming, NM (60
miles from Lordsburg, NM), which has naturally fluoridated drinking water
containing 0.7 mg/L fluoride. The adults were whites of Hispanic or other origin
(74.2% Hispanic); 39.8 years old; 68.2% female; had 12.5 years of education, 27.1
teeth (mean values) (N = 189 examined, which generated 151 subjects). Same water
supply has been in use since early 1900s.
From birth to age 6, and through adulthood, subjects consumed city water (subject age
ranged from 27-65). Some subjects had left town temporarily for military service or
higher education. However, these subjects were included in the study because they
lived in their respective towns for the majority of their life, including the years during
which their teeth were formed.
Adult lifelong residents of Lordsburg or Deming who consumed city water during the
first 6 years of life, and had clearly documented water drinking history through
adulthood. Some subjects had temporarily left town for military service or higher
education.
Based on public records, city water natural fluoride concentrations of 0.7 mg F/L
(Deming) and 3.5 mg F/L (Lordsburg) had been constant since the turn of the century.
Individual intake of fluoride was not measured; study protocol confirmed that the
subjects consumed city water from birth through age six (i.e. did not have private
wells) and through most or all of their adulthood. No information was provided
characterizing other sources of fluoride exposure such as toothpaste, mouth rinses, etc.
The report did not state analytical procedures used to determine city water fluoride
concentrations. No other water quality parameters were reported.
Populations from two nearby towns (60 miles apart) in New Mexico, which were
similar on the basis of income, climate and demographics, but had 5 -fold different
concentrations of natural fluoride in their drinking water, were studied to determine
the influence of water fluoride concentration on teeth (fluorosis and caries). Subjects
selected were -30-60 years old, had consumed city water during their first six years of
life as well as through most or all of their adulthood, and had been born in the
community. Those with a history of private well use were excluded.
Short absences (undefined) during adulthood for military service or education were
acceptable to the study protocol. All eligible adults were recruited (due to small
community size). Authors estimate that 88-90% of those eligible to do so did
participate.
77
January, 2008
-------
PARAMETERS MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
Dental caries
All 32 teeth of each subject were examined for dental fluorosis and caries and scored.
Fluorosis severity was reported using Dean's 1942 classification scheme, except that
the category of "severe" was split into two categories, one used when the pitting was
discrete (severe), and the other when pitting was confluent (very severe). To
calculate the community fluorosis index, both severe and very severe cases were
scored as category 4. The classification for a given person was based on the most
severe fluorosis seen for two or more teeth. Duplicate scoring on 29 people showed
77% concurrence, and 92% were scored within one fluorosis category of each other.
Dental caries were examined using the criteria of Radike (1968) and the DMFT index.
Teeth that had not erupted or were lost to trauma were not scored. Results were
reported for <28 teeth/person and excluded the third molars and teeth with crowns.
Agreement on diagnosis was 94% between examiners for all teeth, and 83% for teeth
with caries (did not state if exams were made in duplicate, as for fluorosis), indicating
reliability.
All teeth of each subject were examined for dental fluorosis and caries (DMFT,
crowns). Fluorosis severity was evaluated using Dean's 1942 classification scheme,
but the qualitative description of severe fluorosis specified if the pitting was discrete
or confluent (latter considered very severe). Dental caries were examined using the
criteria of Radike (1968) and the DMFT index.
The number of decayed teeth, missing teeth, filled teeth, and the DMFT were
compared in subjects of the two towns by the Student t-test or the Mann- Whitney U
test. Comparisons were made for all subjects, and for age subgroups 27-40, 41-50,
and 51-65. In addition, the results were analyzed by covariant linear regression, using
the city of residence as the independent variable, and various dental conditions or the
DMFT index as the dependent variables, with adjustments made for extraneous
effects such as education, ethnic group, gender, age, and dental care.
Residents of Lordsburg had much more severe fluorosis (98.8% moderate to very
severe) than residents of Deming (95.4% normal to very mild), as shown in study
Table 2 (copied from p. 326 of study). The mean community fluorosis index (per
Dean, 1942) was 3.74 for Lordsburg and 0.3 1 for Deming.
Table 2. Number and percent of subjects by city and fluorosis classification
Fiyuioilfc fUwtirtiltiiaj
Qur%NuH- V«n yw
<•'"> NOB.J! iblc -HIM Mild ModTUt torn «.«,,
L«**J<« 1 1 3J 61 (,'.
"~^l !OW* IU£%I P&fK'- ;Sti.W fg?Rl
I>emi«i5 jfl ^3 1? 2 *1
"=15! IMS; llIZS) itISSj |!3"«i |JS!P
The overall DMFT score (per Radike 1968) was 7.0 for Lordsburg and 8.7 for
Deming, suggesting that the Lordsburg residents had better protection against dental
caries; these differences were significant (P = 0.0041). The difference in the DMFT
was largely due to a higher value for the "filled" component, as shown in study Table
3 (copied from p. 326 of study). The DMFT also included rarely encountered
components not shown in the table (e.g. restored teeth with recurrent caries or
fractured enamel), which is why the DMFT score is in all cases greater than the sum
of decayed, missing, and filled teeth. Covariant linear regression analysis indicated
that the Lordsburg residents had 0.09 more decayed teeth, 0.24 fewer missing teeth,
1.71 fewer filled teeth, and 1 .67 fewer DMFT per person than in Deming; only the
differences in filled teeth and DMFT were statistically significant.
Table 3 . Comparison of mean DMFT and selected components by city and age of
lifelong resident adults.
78
January, 2008
-------
Effect of dental fluorosis on
dental caries
STUDY AUTHORS'
CONCLUSIONS:
Decwil M«»«i£ fiUnl OMfT
POI.JJ l« D« 1 D 1. D L St
Aii
I>> = !1» 1X8 Hi ft 14 8,9 MtS 7,0 8»S
KM 40
t'tiilW 01 0.7 14 is 3.6 4,4 "1,9 fi«l
4! in H>
("-S5! 1.5 «'> 8.4 i; 2.4 e.4fj 51 tilt]
-------
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/date:SM
01/03/2007
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY FOR
DOSE-RESPONSE MODELING:
CRITICAL EFFECT(S):
teeth in severe dental fluorosis increases their susceptibility to caries; and (3) high
fluoride levels (such as of 3 .5 mg/L) are a cosmetic, and not a health-oriented,
concern.
The authors acknowledge that is it unknown whether or not the advantage is due to
fluorosis per se or another parameter associated with fluorosis.
Radike, A.W. 1968. Criteria for diagnosis of dental caries. In: Proceedings of the
Conference on the Clinical Testing of Cariostatic Agents. Chicago, American Dental
Association.
This was a well-conducted study from which logical conclusions were drawn. The
demographic profiles of the two study populations were well matched. The study
clearly showed an increased severity of fluorosis in the Lordsburg, NM, subjects, who
were exposed to higher natural fluoride concentrations (3.5 mg F/L vs 0.7 mg F/L in
Deming, NM) in the drinking water. The study was internally consistent in that for
both populations, a given tooth type (molar, premolar, anterior) exhibited a similar
susceptibility to caries, and a greater level of water fluoridation was associated with
and increased resistance to caries.
The primary study weaknesses were (1) a lack of quantitative individual cumulative
exposure data from the drinking water, and from other possible sources such as
toothpaste, mouth rinses, etc. and (2) no evaluation in individuals, but only in
individual teeth, of the relationship between the percent of sound teeth and fluorosis
severity (results of teeth within a given mouth are dependent variables). The study
would have been more powerful if more people and an intermediate fluoride exposure
level had been evaluated.
Based on absence of an increase in caries formation in the presence of fluorosis
(approximately 38% of subject teeth in high-F community ranked as severe/very
severe fluorosis), the NOAEL was >3.5 mg/L lifetime exposure.
NOEL for severe fluorosis is >0.7 mg F/L lifetime drinking water exposure, but <3.5
mg F/L lifetime drinking water exposure.
A LOAEL was not identified. Authors considered the effects of fluorosis on the teeth
to be cosmetic, and that fluoride conferred protection against caries in molars of both
populations (low-dose and high-dose).
Not suitable ( ), Poor ( ), Medium (X), Strong ( )
The critical endpoint of dental fluorosis was clearly shown to be correlated with a
higher intake of fluoride in the drinking water. Although individual exposures were
not quantitated, intake could be estimated by using default or empirical values for
water intake.
Dental fluorosis and caries in molars, premolars and anterior teeth
80
January, 2008
-------
Englander, H.R. and DePaola, P.F. 1979. Enhanced anticaries action from drinking water
containing 5 ppm fluoride. JADA 98: 35-39.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT
ANALYTICAL METHODS:
STUDY DESIGN
Dental caries (permanent teeth)
USA: Prevalence study of dental caries in male and female children aged 12 to 15 years in
seven communities in five states (MA, NC, MI, IL, and TX); dates of the study were not
provided. All children included in the study were lifelong residents of their respective
communities, and none had lived away for more than 30 consecutive days during any
calendar year.
A total of 1,878 white adolescents (aged 12 to 15 years) were examined for caries on
permanent teeth by one of two dentists. The children were examined in Boston, MA (<0. 1
ppm fluoride); Danvers, MA (approximately 1 ppm fluoride); Mecklenburg County, NC
(<0. 1 ppm fluoride from well water); Kalamazoo, MI (approximately 1 ppm fluoride);
Stickley, IL (approximately 1 ppm fluoride); Charlotte, NC (approximately 1 ppm fluoride);
and Midland, TX (5 to 7 ppm fluoride). The mean age of the examined children ranged from
13.3 to 13.5 years depending on the community studied. The dates of the dental
examinations were not provided in the study report. Subject children in each city were
distributed similarly according to age, gender, race and socioeconomics factors. (The
demographic data were not provided in the study report.) All children included in the study
were lifelong residents of their respective communities, and none had lived away for more
than 30 consecutive days during any calendar year. The geographic areas were chosen
because, at the time, they presented few administrative difficulties. Data collected in each
community were based on all the children who volunteered for the examinations.
None.
Children were lifelong residents of the communities so their exposure period was since birth
to the age at which study examination took place (12-15 years' old).
Adolescents (aged 12 to 15 years) were exposed to different levels of fluoride in the water
supply of seven communities in five states: Boston, MA (<0. 1 ppm fluoride); Danvers, MA
(approximately 1 ppm fluoride); Mecklenburg County, NC (<0. 1 ppm fluoride from well
water); Kalamazoo, MI (approximately 1 ppm fluoride); Stickley, IL (approximately 1 ppm
fluoride); Charlotte, NC (approximately 1 ppm fluoride); and Midland, TX (5 to 7 ppm
fluoride). The children were also categorized into three groups; those who consumed either
fluoride-deficient drinking water, fluoridated water containing approximately 1 ppm fluoride
or water containing 5 to 7 pm of naturally occurring fluoride.
Samples of fluoridated water obtained at the time of dental examination showed that the
fluoride concentration in Danvers was 0.67 ppm. The previous high level of fluoride in the
Midland (TX) water (5 to 7 ppm) had been reduced by dilution to 0.3 ppm fluoride ten
months before the dental examinations were conducted because the concentration had
exceeded the MCL of 1.6 ppm fluoride for this area as set by USEPA. Children in
Mecklenburg County NC generally consumed fluoride-deficient water from shallow wells,
but they frequently drank soft drinks and other beverages processed with fluoridated water
from nearby Charlotte, NC; and many subjects also visited Charlotte often, and were
herefore frequently exposed to Charlotte city water. The children from Danvers MA also
drank beverages bottled and canned in nearby Boston MA, and many visited Boston often
(and were assumed to consume "fluoride deficient" Boston city water).
None provided.
The objective of the study was to determine the prevalence of dental caries in adolescents
who, since birth, have consumed either fluoride -deficient drinking water, fluoridated water
containing approximately 1 ppm fluoride or water containing 5 to 7 pm of naturally
81
January, 2008
-------
occurring fluoride. One of two dentists conducted examinations according to the method of
H.T. Dean (personal communication, 1953) on 1,876 adolescents (aged 12 to 15 years)
exposed to different levels of fluoride in the water supply of seven communities in five
states: Boston, MA (<0.1 ppm fluoride); Danvers, MA (approximately 1 ppm fluoride);
Mecklenburg County, NC (<0.1 ppm fluoride from well water); Kalamazoo, MI
(approximately 1 ppm fluoride); Stickley, IL (approximately 1 ppm fluoride); Charlotte, NC
(approximately 1 ppm fluoride); and Midland, TX (5 to 7 ppm fluoride).
PARAMETERS
MONITORED:
Monitored parameters included DMFT and DMFS. One of two dentists conducted
examinations on the permanent teeth to determine the prevalence rate of dental caries in
adolescents (aged 12 to 15 years) in seven communities of five states. All examinations
were conducted according to the method of H.T. Dean (personal communication, 1953) with
the use of mouth mirror, explorer and portable field equipment, including a dental light.
Good agreement was found between the examiners with mean scores for decayed (D),
missing (M), and filled (F) permanent teeth on the same group of 100 children in each of the
communities consistently showing an insignificant difference of less than 0.2 DMFT
between the examiners.
STATISTICAL METHODS:
Since there was good agreement between the examiners, data from all examinations were
pooled for presentation, analysis and interpretation. Differences among means were tested
by analysis of variance; the method of Scheffe was used for multiple comparisons.
RESULTS:
The number of children examined in each community and the caries experiences are
illustrated below.
REVIEWER'S NOTE: Due to the tight binding of the journal, the figure caption ("Dental
caries experience of 1,878 adolescents who consumed water containing varying
concentrations of fluoride") could not be captured in the image below.
Mm
No.
2.0 ,
Wo. OMP
Tttih
yrs.r natita: to »*ea U.S.
eommuntliss will! WmflfBmera
8r Fltmride-tteftsbnt Wamr
MUMBEB
EXAMINED
WiTON, DANS/IRS.
MASS, MASS.
Study results in Tables 1 and 2 are shown directly from Englander, 1979.
82
January, 2008
-------
Table 1 • Distribution of dental caries according to mew numbers of tooth surface
sites.
City
Mean ±SE of
total DMPS*
Pwximal Oeelusal
surfaces surfaces
Bucco lingual
surfaces
Afltcrtoi
surfaces
Boston (PMMietent)
Danvers, Mass (flu»ridated)t
Mecklenburg County,
NC {F-deficienW
Kalamazw, Midi (fluorklated)
Stickney, III (fluoridated)
Charlotte, NC (fluoridated)
Midland. Tex fS to 7 ppm F) ..
Mean difference between «»ch fluoridated city v» Bwwn, tu*i Midland « «v»ry city is significant (P <.0l),
* Proximai. occhwal, mi buoLolingual tooth Burfacws.
+ Analysis of w*Ur stowed 0,6? ppm P.
Table 2 « Distribution of 1,878 children according to dental enries experience (DMFT) to seven eommiinliies
in five states. _™_™™_
± -
8-60 ±
-------
PROFILER'S ESTEM.
NOAEL
PROFILER'S ESTEM.
LOAEL
SUITABILITY FOR DOSE
RESPONSE MODELING
CRITICAL EFFECTS:
to water supplies from nearby communities with quite different fluoride concentrations (e.g.,
Danvers, MA and F-deficient Boston, MA). These observations tend to compromise
interpretation of study results.
The study design did not estimate a NOAEL.
The study design did not estimate a LOAEL.
Not suitable,O; Poor (X); Medium (_); Strong (J
Dental caries (permanent teeth)
84
January, 2008
-------
Ermis, R. B., F. Koray and E.G. Akdeniz. 2003. Dental caries and fluorosis in low- and high-
fluoride areas in Turkey. Restorative Dentistry. Volume 34 (5): 354-360.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
Dental fluorosis and prevalence of dental caries
Cross-sectional survey
Turkey/Izmir and Isparta. 278 (1 14 girls and 164 boys) schoolchildren, 12 to 14 years
old, in Turkey. In the study, 149 children were from the city of Izmir, a low-fluoride
area, and 129 children were from Isparta, a high-fluoride area. All children were
required to have been residents of the study site.
None
Exact duration of exposure was not stated, although children were required to be a
resident of the study site and using the public water supply continuously as their
drinking water source. In addition, they were required to have no nutritional
deficiencies; however, Ermis et al (2003) provide no data or protocol characterizing
how nutritional status was determined. Data were collected in the Spring of 1999.
| ^^^^^1
Fluoride Girfs Boys Total
Area . concentration (ppm) (n = 114) (a=(64) (n = £78) Mean age SO
LFA 0.30-0,40 6€ 83 t« 13,03 080
HFA1 1.42-1,54 27 36 63 12,83 083
HFA2 t. 5M.66 at 45 66 13.11 0.83
JSO s siimfam deviation; LFA = few-Ouonda area; HFA1 w KgMluwkta area 1; HW2 n Msh-tluerae »« a.
Table 1 was copied directly from the study report (Ermis et al., 2003) indicating the
gender, number and ages of the participants and the levels of fluoride concentration
(ppm) of the water supply.
PROFILER'S NOTE: The fluoride in the water supply occurred naturally as the water
supplies in the study communities had not been deliberately fluoridated.
Two examiners were used for the study. Prior to the study, the two examiners were
involved in a 2-day training session to standardize how they rated the incidence of
fluorosis and caries. At the end of the training, the two agreed on the scoring of
fluorosis and caries greater than 90% of the time. Intra-examiner reliability was
performed on 10% of the sample, and the results showed an agreement of 77% to 94%
for scoring of the fluorosis and caries. The examiners used the Tooth Surface Index of
Fluorosis (TSIF) by Horowitz et al. (1984) for scoring the fluorosis. Caries were
diagnosed according to the World Health Organization (WHO, 1997)
recommendations. Examinations were performed under natural daylight using plane
mouth mirrors and explorer tools.
The method of measuring the fluoride in the water supply was not provided. The cities
of testing were Izmir, which was considered an area of low-fluoride and had a fluoride
concentration ranging from 0.30 to 0.40 ppm in the drinking water; and Isparta, which
was a high-fluoride area in which fluoride ranged from 1.42 to 1.66 ppm Both cities
did not have fluoridation programs and the fluoride occurred naturally. Values of the
fluoride concentrations were obtained by the Ministry of Health, Public Health
Laboratories.
The study subjects included 278 (1 14 girls and 164 boys) schoolchildren, 12 to 14 years
old, in 2 Turkish cities. The subject population included 149 children from the city of
Izmir, a low-fluoride area; and 129 children from Isparta, a high-fluoride area.
85
January, 2008
-------
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
To initially select the participants in the study, four schools in both cities were selected
by random sampling from a list of secondary schools. Then, 12-14 year old
schoolchildren were randomly selected from each school.
Clinical dental examination of each child was performed once (Spring 1999) by two
examiners using natural daylight, plane mouth mirrors and explorer's tools.
The examinations scored dental fluorosis (Tooth Surface Index of Fluorosis) and caries
(WHO criteria). Children were also asked about their tooth-brushing frequency.
Fluorosis was assessed by using the Tooth Surface Index of Fluorosis (TSIF) (Horowitz
et al., 1984) which is defined in the Section 2 of the report. Individuals having at least 2
different teeth with a TSIF > 1 were defined as a case of fluorosis.
Caries was diagnosed according to the World Health Organization (WHO, 1997)
recommendations. Diagnosis was made when a lesion in a pit or fissure or on a free
smooth surface had a detectable softened area, or there was a temporary restoration. A
separate score was given to each facial and lingual surface of anterior teeth and to each
facial, lingual, and occlusal surface of posterior teeth. Indices used for evaluation of the
caries were: (1) decayed, missing and filled permanent teeth (DMFT) and (2) decayed,
missing and filled permanent surfaces (DMFS). Radiographs were not taken on any of
the children.
Children were also questioned about the frequency of their tooth-brushing and allowed
the following answers: didn't brush, brushed once a day, more than once a day, or
brushed irregularly.
No information about exposure to any other sources of fluoride or the amount of water
consumed was included in the study protocol description.
TSIF fluorosis scores were statistically analyzed using the correlation analysis to
determine any differences between fluoride exposure groups. Kruskal-Wallis and
Mann- Whitney U test were used to compare the tooth-brushing frequencies between the
low- and high-fluoride areas; the caries indices were compared to the tooth-brushing
frequencies. The tooth-brushing frequency efficiency in DMFT and DMFS indices was
evaluated using analysis of covariance. Spearman's correlation analysis was used to
determine the association between severity of dental fluorosis and caries prevalence.
The level of significance was p< 0.05.
Children in the low fluoride areas had no evidence of fluorosis. Table 3 is copied
directly from Ermis et al. (2003) and indicates the TSIF scores for all the children
examined in the high-fluoride areas. The scores indicated that the prevalence and
severity of fluorosis increased as the fluoride levels increased. Figure 2 was also
provided from Ermis et al. (2003) and indicated the cumulative distribution of the
average TSIF score per child in the different fluoride groups. According to the report,
the percentage of children with TSIF scores > 1 was 29% and 77% in the HF area 1 and
HF area 2, respectively. According to the TSIF scoring index employed by Ermis, et al,
pitting occurs at TSIF > 5, with initiation of brown staining at TSIF = 4.
• TABLE 3 Percent distribution of TSIF scores for all permanent tooth ^^^^^^^H
surfaces according to high-fluoride areas 1 and 2 ^^^^^^H
No. of Distribution of TSIF scores (%)
High-fluoride area surfaces 01 234567
1 (1.42-1. 54 ppm) 3,910 32.38 41.56 6.34 4,78 14.07 0.56 0,28 0.05
2 (1.55-1. 66 ppm) 4,266 14.58 29.08 11.20 13.69 25.06 3,31 1,69 0.49
TSIF = Tooth Surface Inttex of Fluoreeis,
86
January, 2008
-------
I
S
o
B
o
I
£D
e
£
100
801
50
40
-•-1.55-1.66 ppm fluoride
130.00 -m-i .42^1,14 ,pj)m. fluoride
77,27
SfO.
95
\ \
20
0
\2S.
^
.57
,
-13.64
15.87 "^^^S-lT^I .52 0.00
0 0.5 1,0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Average TSIF score
Fig 2 Cumulative distribution of TSIf- fluorosis prevalence and
severity for different fluoride-exposure groups,
This figure is interesting because it does show dose response relative to the severity of
fluorosis in two places where the differences in fluoride are small. It is unfortunate that
the study authors did not provide information on differences regarding other sources of
fluoride or water intake that could be examined.
PROFILER'S NOTE: The profiler agrees with the study author that the prevalence and
severity of fluorosis increased in the area with the higher amount of fluoride. However,
the number of children with adverse TSIF scores (> 5, considered severe, was 0.89% in
HF area 1 and 5.49% in HF area 2) were actually fairly low in both areas. The low-
fluoride data in a comparable format should be provided even though it was stated that
no fluorosis was identified. Further breakdown into gender was not needed but could
have been helpful in determining any trends. Information on the methods used to test
fluoride concentrations in the water and the frequency of the water testing was not
included. The captions for Figure 2 do not readily indicate that the results illustrated are
for high-fluoride communities only and that the X-axis unit is the average TSIF score
per child; further, the distribution of TSIF scores/child cannot be determined from the
high-fluoride community TSIF distribution (Table 3) presented.
Dental caries
Table 5 is copied directly from Ermis et al. (2003) to indicate the caries data in relation
to the TSIF score. Another table, Table 4, was not copied as it just provided caries data.
Ermis et al. (2003) states that the findings in the study indicate no significant difference
in the prevalence of caries prevalence in the permanent teeth of children exposed to
0.30-0.40 ppm and 1.42-1.66 ppm fluoride in the drinking water. However, the
relationship between DMFT and DMFS scores and the frequency of tooth-brushing was
significant.
87
January, 2008
-------
FbO'iLis sxpos j e
.-PA
UFA !
"-FA2
~bc jnorus1" iny <
D:i "Ot btt s'i
''-M>. %1v
Qrw.idnv
Merc 3" o onct-.iday
TSIF bCO'c
149
63
22
49
115
92
105
0,84
' 3C
1.59
1.61
1.10
0.53
1 46
1 4'
1.44
1.46
1.21
0.80
1.5b
1 78
2.45
2,65
0.85
2 24
252
2C-J
2.91 I
2.94
2.38
1.58
29 ' 4t> * a? ?,.-
i - PO > it-doe , 3V1Ff- tt/il » r v.J ru-n and 'liar, tcot SC -
I.!.,' <.,'d OCMtun QMF; *ct.i <"> i»Pd rrssnr. an- ilM "i>f«i.t.t,
i-fA low fluj-ide ate?, HcAi ~ 1 >ai lldOidsarea' hFAi _ Urt
PROFILER'S NOTE: The profiler agrees that there was a lower DMFT and DMFS in
the low-fluoride area compared to the high-fluoride areas but there was not much of a
change between the two high-fluoride areas. The profiler also agrees that tooth brushing
frequency does appear to help with the incidence of caries; however, without additional
data, the profiler can not make any definite conclusions. The study author states that the
lower fluoride area children had better oral hygiene (i.e. more frequent tooth brushing)
but one cannot make that distinction from the data provided. Also, the
instruments/cleaners used for tooth brushing were not characterized.
STUDY AUTHORS'
CONCLUSIONS:
Findings of the study indicated that the significant difference in prevalence and severity
of dental fluorosis was closely associated with fluoride concentration in the drinking
water. While there was no significant difference found between the caries prevalence
and fluoride concentration, poor oral hygiene had a significant effect on the higher risk
of dental caries in the high-fluoride area.
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
World Health Organization (WHO). 1997. Method of assessing dental caries. Oral
Health Surveys: Basic Methods, ed. 4, Geneva, Switzerland, p. 39-44.
PROFILER'S
REMARKS
Initials/date
DFG/11-06
12/14/06
Adequate data were presented but the profiler could not determine how some
conclusions were made with the data presented. Some of the percentages included in
the paper were not consistent with the data provided (See Profiler's Note under Results
section.). Also, there was a lack of confirmation about the subject's duration of
exposure to the drinking water and if there was exposure to other sources of fluoride
(i.e. supplements, food). Additional information on how the fluoride levels in the water
were measured (i.e. analytical techniques) and how often the levels were measured to
account for groundwater fluctuations would have added value to the study.
The study did try to minimize subjective differences by only having two examiners and
having a "training" session with these two individuals to allow them to become
consistent on their scoring techniques. The use of the TSIF scoring and the indices of
dental caries by the WHO also helped in the study objectiveness. Statistical analysis
appeared to be adequate although more data to support the conclusions are needed. The
addition of the tooth brushing frequency added another variable that took away some of
the study's objectiveness. The study did not address how tooth brushing was performed
and if it involved the use of toothpaste containing NaF as most commonly do. Also,
from the data provided, the profiler could not distinguish which children were the ones
with the better oral hygiene.
January, 2008
-------
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
Study design was not suitable for development of a NO AEL.
Study design was not suitable for development of a LOAEL.
Not suitable (J, Poor (X), Medium (J, Strong (J
Fluorosis and prevalence of dental caries
89
January, 2008
-------
Forsman, B. 1974. Dental fluorosis and caries experience in high-fluoride districts in Sweden.
Community Dent. Oral Epidemiol. 2:132-148.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL
POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL
METHODS:
Dental fluorosis; caries
Cohort
Sweden/ Gadderas (pop. 90): 39 individuals aged 2-35 years; 28 were born in Gadderas and 15
were less than 15 years old.
Sweden/Paskallavik (pop. 900): 190 schoolchildren, born in the years 1955-1966, 61 pupils were
born and raised in the area.
Sweden/Billesholm (pop. 3000): 300 schoolchildren; 133 were born and raised in the district.
All children received regular dental care at the district clinics from the time they were either 3 or
6 years of age.
Sweden/Eskilstuna (pop. 93,000) and Kronoberg county (pop. 168,000): Schoolchildren (data
for 160 children from Kronoberg county were used in comparison studies).
No consistently set exposure periods existed. In Gadderas, exposure at ~ 10 ppm ranged up to a
maximum of 27 years (from 1946 when a new water source was implemented to 1973, the year
this study was submitted), with exposure initiating from birth or at intervals up to 14 years old.
In Paskallavik, exposure at ~10 ppm ranged up to 9 years and was categorized into periods less
than or more than 4 years. In Billesholm, exposure at ~ 5 ppm ranged up to 12 years.
The following exposure group designations were used in the study analysis:
Gadderas: Starting in 1946, homes in the town were connected to a new water source and all
homes were connected by 1950; mean fluoride concentration from 1969 to 1973 was 10.1 mg/1.
For the purposes of this study, the fluoride level was considered to be ~ 10 ppm.
Paskallavik: The water source had a fluoride level of 7-10 mg/1 from mid-1956 to beginning
1965; prior to 1956, private wells with low fluoride content were used, and after 1965 the water
source was changed, with a fluoride content of 2.0-2.5 mg/1. For the purposes of this study, the
fluoride level was considered to be ~ 10 ppm.
Billesholm: Water was obtained from two deep wells; from 1957 to 1969 fluoride level varied
between 4 and 7 mg/1, but mostly was around 5.5 mg/1. From 1969 to 1973 the fluoride content
was 1-3 mg/1 or less. For purposes of this study, the fluoride level was considered to be ~ 5
ppm.
Eskilstuna had a water fluoride level of 1.2 mg/1; the same ground filtration system has been in
use since 1913. Two districts in Kronoberg county have 0.9-1.0 mg F/l in the water and the
other two districts, 1.3-1.7 mg/1, at least since 1950. For the purposes of this study, the fluoride
level was considered to be ~ 1 ppm.
Factors influencing fluoride intake (e.g., diet, fluoridated dentifrices, etc.), other than drinking
water, were not considered in the study report.
Water, enamel/dentin, saliva, and breast milk fluoride data were analyzed using the Orion F"
electrode. Water quality data were not included in the report. For teeth, the crowns were
divided into an occlusal and a cervical half; enamel was separated from dentinby the floatation
method of Manly and Hodge (1939). Fluoride content was determined in buffered solutions of
the washed and dried enamel/dentin powder. After an overnight fast, parotid saliva was collected
with Lashley cups under stimulation with 6% citric acid on the dorsum of the tongue. Breast
milk samples were taken for three days in the hospitaljocated in the county capital, where water
90
January, 2008
-------
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL
METHODS:
RESULTS:
Dental fluorosis
fluoride level was <0.2 mg/1. After some weeks at home in Gadderas, samples were taken for 2
days: a morning sample after fasting and an afternoon sample, hours after ingestion of -0.5 1 of
Gadderas water.
The study included residents from communities in southern Sweden, mostly school children,
exposed to fluoride at levels of either ~1 ppm (n= not reported, Eskilstuna and Kronoberg
county), ~ 5 ppm (n=300, Billesholm) or ~10 ppm (n=229, Gadderas and Paskallavik). Water
fluoride data were often checked by personal analysis.
The occurrence of fluorosis and caries in primary and permanent teeth was investigated either by
examination or dental records, using Dean's index and DMFT or DMFS scores, respectively.
Caries generally was obtained from dental records; in subjects from Gadderas whose primary
teeth had not already exfoliated, caries registration was done at the same time as the fluorosis
examination. Color photographs of typical cases were taken in every district. Examination
details were not reported, including location where examinations occurred, lighting conditions,
or equipment used. Case histories, including water supply data, places of birth, migrations, etc.,
were collected from different sources: waterworks records, school authorities, the vital statistics
system, etc. Several data from Gadderas and Billesholm were collected from previous surveys.
Factors influencing the incidence and degree of fluorosis were evaluated, including duration of
breastfeeding and prenatal fluoride exposure. Information on feeding during the first year of life
was obtained by questionnaire; a second questionnaire was used by mothers in Paskallavik and
Billesholm to assess fluoride exposure before and during pregnancy, in addition to during of
breastfeeding.
The relationship between fluoride exposure and content in enamel and dentin was analysed, as
well as the relationship between plasma and saliva fluoride. Fluoride content was determined in
exfoliated primary and permanent teeth extracted for orthodontic reasons from individuals in the
~10 ppm and ~ 5 ppm areas. Saliva samples were taken from 21 subjects. Fluoride content in
breast milk was determined from two patients residing in Gadderas for less than 2 years.
Dental fluorosis was recorded according to Dean's index; fluorosis was scaled as none (0),
questionable (0.5), very mild (1), mild (2), moderate (3), or severe (4). Caries generally was
obtained from dental records and covered both DMFT (decayed, missing, and filled permanent
teeth) and DMFS (decayed, missing, and filled permanent tooth surfaces).
Statistical methods were not detailed in the study report. From figure legends and brief text, it
was concluded that regression analyses were used. Chi-square test was used to determine
significance inbivariable analyses.
Figures and Tables were copied directly from Foreman (1974). Figure 2 summarizes the percent
distribution of fluorosis in permanent teeth according to severity and community; data from
primary teeth is not shown. At ~10 ppm, all individuals born in Gadderas after 1950 had
moderate to severe fluorosis in all permanent teeth. Most primary teeth had moderate to severe
fluorosis for canines and molars and up to mild fluorosis for incisors. Children who moved to
the district after the age of 18 months showed no fluorosis in the primary dentition. Severe
fluorosis also was noted in Paskallavik. All but one child born between 1957 and 1961
(exposure > 4 years) had moderate to severe fluorosis; 27% (7/26) had severe fluorosis on all
teeth. Children born 1962-1964 and who had less than 4 years of high fluoride exposure showed
milder fluorosis (only 40% with moderate to severe fluorosis). There were 1 1 children born
within the 1.5 years after the fluoride content was lowered to ~2 ppm (in 1965); 9 of the 1 1 had
fluorosis of grade 1 -4 in their primary teeth. Thus, a long period of exposure is necessary, even
with very high water fluoride concentrations, for a severe degree of fluorosis to occur in
permanent teeth (difference between exposure durations was significant at pO.Ol). In primary
teeth, degree of fluorosis was generally moderate to severe.
At the ~5 ppm fluoride level, 28% showed moderate to severe fluorosis in the permanent teeth,
91
January, 2008
-------
50% mild fluorosis, 25% very mild, 1.5% (2/133) had no fluorosis. In the primary teeth, all
degrees of fluorosis were observed, generally mild; 20% (of n=67) had no fluorosis in the
primary dentition.
Gender, duration of breast feeding, and prenatal exposure to fluoride were evaluated for their
influence on the occurrence and degree of fluorosis. Table 2 shows that, there was a tendency
for more severe fluorosis to occur in boys compared to girls at ~5 ppm fluoride (Chi-square test,
p<0.05), but the sample size was small.
Talilr ?. C.uni[j:t!j.»Mi Lrinrrn bow and ((iilv fkRiff" of fly
urotit in |j!i'tiio*3« .iitd v«'Mid mr.Urv •>, ) fsjuii atr.i
Pt:i
distnbosian nf l)m\
uTirdiv? to de«'<«' of
Bo i*
Girls
39
40
"ie,
Figures 9-12 show the occurrence of fluorosis in permanent incisors and molars in relationship
to duration of breastfeeding (up to 10-12 months) in areas with different water fluoride levels
(data for primary teeth are not shown). Figure 12 shows that in the -10 ppm areas, severe
fluorosis occurred with a long duration of breastfeeding. Fluorosis in the primary dentition was
milder in children breastfed for a longer period. In the other districts (Figures 9-11),
increasingly severe fluorosis occurred with higher water fluoride levels for the same number of
months of breastfeeding.
FLUOROSIS
INDEX
0.5
0
0 5 10
MONTHS ON
BREASTMILK
F%» 0. Rrlwiwslup b«asffe«Hng - enamel flmsrom of
jwiranrat incUors awd firsl molacs, ^ 1 i>pm ai'ea, CL9-
1.0 mf F/l,
FLUOROSIS
INDEX
1
0.5
0
Fif, 10, RelJ
perawemt i
1.7 m( F/t,
5 10
MONTHS ON
BREASTMILK
br«wd«diog - epamd flyoitssis ei
K| flrM t»iare» ^, I ppm area, 1.3™
92
January, 2008
-------
FLUOROSIS
INDEX ft"
% - - |j *', • f - • I
3 - : •
3
2 . !.-'•,•• .If,--..
. •„,.. ......... 2
1 - . - .,,--......,. t
* " •..-•.-.'• |
0.5 - • 03
o (- . . o
OROSIS
:x
-•»••;•;; •
0 5 10 0 5 10
MONTHS ON MONTHS ON
BREASTMILK BREASTMiiK
F«. 11- RcUtiomhip brewfmlin, - enamel fluoraii ol F'«?' I?' R*<»>«Wl> b.c«lffdi»g - «at«) fliw«i« ci
ptnn»B«I incisms «id linl molar.", ^ 5 pp,» >r». |«I«I..IWIK mows swl f.rtt molars, -x, 10 |>pn> arcs.
Table 3 summarizes the exposure of mothers to fluoride in the ~5 ppm area before and during
pregnancy, combined with fluorosis data from primary teeth. The degree of fluorosis was higher
in children whose mothers spent the pregnancy in a high (~5 ppm) fluoride area (75% with mild
to severe fluorosis in ~5 ppm area vs. 50% in low fluoride area) and/or lived in a high fluoride
district before pregnancy (85% mild to severe fluorosis with high exposure vs. 41% mild (0%
moderate to severe) with low exposure) (Chi-square, pO.Ol and 0.001, respectively).
r»hlr ,\ J)»gr<* of flucrout 111 cJuldrrit't pfiut,ny t«fth in
jrf,it,in >u \, t \putiYi n{ mc>tiif»\ 10 f di.fMig pregnant1),
titxdk ^C 1 st^aO^ ^v* "? p|^st «i*rHi
A,
IrrKMlHjr clntiibuttoit
titan n jiStgH.ifirt ' t<) «'fprcr of UnrMtms
f> i 5 i-4
•v't jipiti aa,, 40 5 ',:1 v )8
Low r aira 16 "14 8 J! t9
II
l^t JCn»s,1gr 4ti^lfflljii5a' IT
U ! 3 i-4
II T™| IT1 ^ 'J ^ *f *'! C "
**"§»! 1* ^^I^W^fC ^T ~* * J 3l^ « _^
6igfliluani at (he 0 1 ff Icvrl
* Agr > 15 )*,irs, of wtikti .11 ttast 4 vt'ws Bjirnf in -v, 5
p{ tl» .irrj G1 Jt lr«l 1C )»nii iB -N , i Jljur, IrftS.
" Age •<,(!» ^rsrs, Iswitt^ in .« low !•' jrua
Three cases of 'delayed fluorosis' were found. One patient born in a ~10 ppm area moved to a
district with a fluoride level <0.2 mg/1 at 3.25 years of age and at age 17 years had moderate to
severe fluorosis in all teeth except the second molars, which showed mild fluorosis. Two other
patients, born in 1953, moved from a ~10 ppm area to a low fluoride area at age 7 years and had
moderate to severe fluorosis in all teeth, but the third molars which had not yet erupted.
PROFILER'S NOTE: The profiler agrees that the occurrence of fluorosis was greater in
subjects with a longer (>4 years) duration of high fluoride exposure (~10 ppm) compared to
subjects with a shorter (<4 years) duration of high fluoride exposure or with mid-level (~5 ppm)
93
January, 2008
-------
Caries
fluoride exposure. However, it should be noted that the sample sizes were small in the ~10 ppm
areas (n=l 1-26) and much different from that in the ~5 ppm area (n=127). Both the water
fluoride level and the duration of breastfeeding impacted the occurrence and severity of
fluorosis. Further, the mother's pre- and during pregnancy fluoride exposure seems to influence
the severity of fluorosis in offspring, but the sample sizes were small and accurate history of
actual fluoride exposure is difficult to confirm. A definitive conclusion regarding the cases of
'delayed fluorosis' can not be determined since the data were limited to only 3 cases and the two
photographs included in the paper did not reproduce well.
In the ~10 ppm areas, the caries frequency in permanent teeth was unexpectedly high and
extremely high in subjects in Gadderas who were over 20 year old; only one adult was caries-
free. The caries frequency in primary teeth was low; in Gadderas, caries in primary teeth
occurred only in children who moved into the area after 18 months of age. In the ~5 ppm area,
the caries frequency in permanent teeth was less, except for about 20 teenagers, all of whom had
>20 DMFS (data not shown).
Previous investigations for caries and fluorosis with 160 children from ~1 ppm areas in
Kronoberg county showed low caries frequency and, at most, mild fluorosis. Figure 14A
demonstrates the comparison between this control data and data from individuals showing
moderate to severe fluorosis following long (>4 years) and high (~10 ppm) fluoride exposure.
There was a direct correlation between degree of fluorosis and caries frequency, as well as a
direct correlation between age and DMFS count. The regression coefficients for both lines were
not statistically significantly different from one another, but the difference in height between the
lines (5.89) was significant (p<0.01). In the ~5 ppm area, the degree of fluorosis varied greatly
among subjects so comparison for caries frequency was made between those with no to very
mild (<2) fluorosis and with moderate to severe (>3) fluorosis. The results are presented in
Figure 14B. The higher degree of fluorosis was associated with higher caries frequency
(pO.Ol).
DMFS
20
15
10
5 ,
DEAN'S INDEX-3
y:t.408x-8.783
s= 9.39
l ppm
DEAN'S INDEX «£ 2
fiO.778x-il.53
10
15
20 25
A6E, YEARS
DEAN'S INDCXfe 3
y* 2,2iSx-16,74
ssA.92
BEAN'S INDEX *
y= 1.326x-8.96
5 = 2.14
10 tS
A8E, YEARS
94
January, 2008
-------
Fluoride content in
enamel/dentin and saliva
Fig, 14. Age - rarirs rcgrfsiitwi iiiifs for groups wifh differ-
ent (Srgm* of t\ ,«««?
the r«{»eclJve ngc
ftoMmh Tin- fvgreswo
n ljura ciivtl1
raij^ef, and ilif <;aries avrr»g<*» ai'«*
marked by point! on the )mr»
s denotes standard deviations of the resfjccsiyr
A n^ 10 ppm, UEAH'I
Numbers of obtervatio
B. DEAN'S index ^ 3
ltn«
io'iejt J\ 3 vcnuf r\* I ppm, DKAN'"!
i: -v, 10 ppm: 38, ^ 1 ppi«r 160,
venn*4^,2, toil) in *^ A ppn« at«ra
Numbeis of nb$f rvadotit: 1>. i ^>
S3:3?, D, i,
^2:91
Figure 15 shows a comparison between deft count (decayed, extracted, and filled teeth) at 6-7
years in areas with different fluoride levels. Increased water fluoride level significantly reduced
caries frequency in the primary dentition even at ~10 ppm. No association between high degree
of fluorosis and high caries frequency was demonstrated in the primary dentition.
10 • T
t - ;
8. r.
t i
i • ;
T
'li
4 .
3 ,
2 .
i . :
I
-1 0
c*»1
Fij, 15 Cariet in
J
1
__L
i 1
P
!
*
! ,
2-8
Hr-j
5
I
f__ t
Oppm o»bppm cK"lpp.m
the psimiry denlrtton oi thilrffeo born
and reared in areas with different water r content.
'\, JO ppm » Piskallavik
-1 "— children
ppm, a ~ 5
3 - children
n - 6
I w chilslrf n
p|3i% n ™ 7
2«B = clilMrei
b&ro Ihe y^ar beloi? i*iuon
hor
-, k
i tlie year «if
B lh«- ytai
introdiicirif >
v 10 ppm,
after introdLcinf »\» 18
ppw, n — 40
*._5 ppm ~ Bitldboltn, n — 1^4
'v I ppm = ilkiljtuni, n = 2,170
Pittallavik 2-8 children and ftillesholm chiWrcr» siynlfi-
camly differmi from
Eskiljiuna rhrldrcis (F <0,l %j.
PROFILER'S NOTE: The profiler agrees that higher degree of fluorosis was associated with
higher caries frequency in both the ~5 ppm and ~ 10
ppm areas;
caries frequency also tended to
increase with age. Data for caries frequency was not included in the report, so results could not
be confirmed.
The mean fluoride level in premolars (permanent teeth, n=3) from subjects in the ~ 5 ppm area
was 447.2 ppm in enamel and 617.1 ppm in dentin, about 4 times greater than the values
obtained from the ~1 ppm area. In the incisors (primary teeth, n=4) of children from the ~ 5
95
January, 2008
-------
ppm area, fluoride deposits in enamel (305.5 ppm in occlusal tissue, 582. 1 ppm in cervical
tissue) and dentin (688.2 ppm and 1200.2 ppm, respectively) also were high. For children who
moved into the area, fluoride content was about 1A of that of the children born there and about
the same as for children in -
Figure 16 (A and B) shows
~1 ppm areas (data not shown).
fluoride deposition in the enamel and dentin of primary teeth in
relation to fluoride exposure in the ~ 10 ppm areas, calculated as the number of post-
mineralization months without breastfeeding. Fluoride level was high in teeth with zero months
of exposure and some increase occurred with increased number of months. The correlation
between estimated fluoride exposure and fluoride content of the dental tissues was strongest for
occlusal tissue (Figure 16A, enamel r=0.73, dentin r=0.66, significant, but level of significance
not indicated), but weak for cervical tissue (Figure 16B, r=0.32 and 0. 18, respectively).
PARTS F/ 108
1800 -
1500 •
1000 -
= 0
500 .
fl
*
*
t
LL.JII|HJ.I.IH. it"in: 1' TT^-i— wi
0
A
Ftg i ij. Relationship F-CAJ>QSSI
PARTS F/108
a
1800
0
a
0 1000
o
500
*
5 10
MONTHS OF
HIGH F -EXPOSURE
0
o o
*
0
o
*
a a
*
a
# *
0 °
*
» *
»
0 5 10
B MOMTHS OF
H1©H F-EXPQSU&E
re - F*iwnc-ni in (kcidaus enamel and clots m. *x* 10 ppffl K«ms, F-es|»surc » smmk»«r ©i'
pffl*!(ialal mmerahzstsoo months wtshnul hte^sffediag A, welusai tmye. B> cwvkal (issue, * enamel, D clrwm.
The relationship between saliva and plasma fluoride levels was plotted in Figure 17 (r=0.40); the
values for fluoride in saliva were lower than those for plasma.
PAROTlC SALIVA f.
MS/i
0.1S
0.10
0.05
Fig,
O
o
* &
©
•
*
005 010
a
6
8
0 »5 020
PLASMA fMG/i
!?. Rflatioi»Iiip plasma F" -• parotid taliva F. • €
t8-Yfar-oM suhjtrti O 35 83 yrir-u
Id stibjffts
PROFILER'S NOTE: Data for fluoride content in enamel and dentin was not included in the
report, so results could not be confirmed. The profiler agrees that there was a correlation
96
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE
NOT FOUND IN NRC
(2006)
PROFILER' Initials/dat
between estimated fluoride exposure and fluoride content of the occlusal dental tissues. Data for
fluoride content in saliva alone was not reported. Results for fluoride content in breast milk are
not presented in the profile since data was of limited value, from only 2 subjects from the ~10
ppm area and inconsistent with each other.
In areas with ~ 10 ppm fluoride, virtually all subjects had moderate to severe fluorosis. In
Paskallavik (~ 10 ppm), a greater percentage of moderate to severe fluorosis was observed in
children with more than 4 years of fluoride exposure. The milder fluorosis seen in children with
less than 4 years of fluoride exposure, due to a change in water supply to lower fluoride content,
may be explained by the fact that the most recently formed layer of enamel was more completely
mineralized because of the lower fluoride supply.
The tendency for boys to in the ~ 5 ppm area to have higher degree of fluorosis compared to
girls might be due to a larger intake of fluoride via water consumption by boys since they have a
greater average body weight and are generally more physically active.Jn low F areas (including
the ~1 ppm population studied here), the extra dosage may not be large enough to be of
importance for incidence of fluorosis, and in the ~10 ppm areas the intake for both boys and
girls is already so high that severe fluorosis results. In the ~5 ppm area, however, even a small
increase in water consumption could weigh the balance.
In the ~ 5 ppm area, and even more so in the ~10 ppm areas, both high fluoride deposition and
severe fluorosis occurred in the permanent teeth even with long periods of breastfeeding. In the
primary teeth, fluoride deposition was high but fluorosis milder. Thus, breastfeeding is
important for the development of fluorosis and for the deposition of fluoride in the dental tissues.
In permanent teeth, there was an increased frequency of caries in the ~ 10 ppm area, also partly
noted in the ~ 5 ppm area; additionally, increased caries frequency was directly associated with
more severe fluorosis. A reasonable hypothesis may be that teeth with severe fluorosis have
hypoplastic enamel, allowing penetration of acid, sugar, and possibly bacteria such that the
increased fluoride content is not sufficient to resist dissolution by acid. In primary teeth, caries
frequency appears to decrease with higher fluoride level in the water supply, despite that fact
that the degree of fluorosis in the ~ 10 ppm areas also was high in these teeth. The pattern of
hypomineralization and/or the pattern of normal mineralization may be different from the
permanent teeth.
The successive improvement in deft count in the ~10 ppm area, Paskallavik, at the time of and
immediately after the increase in fluoride level in the drinking water indicates that a combination
of pre- and post-natal deposition may have the greatest effect.
The high fluoride content in the early mineralized occlusal parts of enamel and dentin in primary
teeth, even with long duration of breastfeeding, indicates prenatal deposition. Further indication
is the relationship between the fluoride exposure of the mother before and during pregnancy and
the degree of fluorosis of the child. The result must be assessed carefully since it is based on
case history information.
The significance of the relatively modest increase of fluoride content in saliva is somewhat
obscure but calculations of the degree of saturation of fluorapatite in saliva indicate that even
this small increase could have a cariostatic effect.
In several aspects, the material presented here is too small to permit definite conclusions,
particularly in regard to Gadderas.
Deft= decayed, extracted, and filled teeth
Manly, R.S. and Hodge, H.C. 1939. Density and reflective index studies of dental hard tissues. I.
Methods for separation and determination of purity. J. Dent. Res. 18: 133-141.
Overall, the study was well-conducted and had adequate study design. The emphasis was on the
97
January, 2008
-------
s
REMARKS
e
SJG/
3/23/07
PROFILER'S ESTEM.
NOEL/ NOAEL
PROFILER'S ESTEM.
LOEL/ LOAEL
POTENTIAL
SUITABILITY FOR
DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
prevalence and severity of dental fluorosis and caries in individuals, mostly school children, in
two areas with high fluoride in the water supply (~10 ppm or ~5 ppm) compared to areas with
low (1 ppm) fluoride concentration. Gender, duration of breast feeding, and prenatal exposure to
fluoride were evaluated for their influence on the occurrence and degree of fluorosis. Results
from the study suggest that both fluoride level and exposure duration influence the prevalence
and severity of fluorosis; further, duration of breastfeeding also influences fluorosis
development. Increased caries frequency was directly associated with more severe fluorosis (in
permanent teeth).
Limitations of the study include:
Data for caries frequency and for fluoride content in enamel, dentin and saliva was not included
in the report, so results could not be confirmed. Results should be interpreted with caution since
sample sizes were generally small, particularly in Gadderas (~10 ppm area) and some results are
based on case history information (e.g., prenatal fluoride exposure), which is difficult to confirm.
Further, the possible significance of drinking water factors other than high fluoride content for
occurrence and degree of fluorosis is not known (e.g., water from the ~10 ppm areas contained
high iron and manganese levels). Finally, other sources of fluoride intake are not considered
(i.e., diet, fluoridated dentifrices).
Based on the prevalence of moderate to severe fluorosis. the estimated LOAEL is 1 ppm
fluoride (1 mg/L) in the drinking water.
Based on the prevalence of moderate to severe fluorosis. the estimated LOAEL is 5 ppm
fluoride (5 mg/L) in the drinking water.
Not suitable (_), Poor (x), Medium (_), Strong (_)
This was a well-conducted and designed study, with some important imitations, including small
sample sizes. The study indicated that moderate to severe fluorosis occurred, generally in a
dose-response manner, in subjects from ~5 ppm and ~ 10 ppm areas, but not in subjects from ~1
ppm areas. A longer duration of high fluoride exposure (e.g., >4 years) increased the prevalence
of moderate to severe fluorosis, and a longer duration of breastfeeding reduced, but did not
eliminate, the prevalence of fluorosis. Although increased caries frequency was directly
associated with more severe fluorosis in permanent teeth, dose-response information for caries
could not be determined from the study since stand alone data was not presented. The study did
not address any issues of plaque or gingivitis.
Dental fluorosis prevalence and severity and caries
98
January, 2008
-------
Franzman, M.R., S.M. Levy, J.J. Warren, and B. Broffitt. 2006. Fluoride dentifrice ingestion
and fluorosis of the permanent incisors. JADA 137:645-52.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL
POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT
ANALYTICAL
METHODS:
Dental fluorosis (permanent incisors)
Cohort study
PROFILER'S NOTE: Franzman et al. used data from Iowa Fluoride Study (IPS); see also
Hong et al 2006a and Hong et al, 2006b.
US/Iowa: Children (343 ; 49% males and 51% females) aged birth to 36 months which were
originally part of the Iowa Fluoride Study was used for this study. Franzman et al. (2006)
used data about fluoride exposure at ages 16, 24 and 36 months and the data obtained from
the dental examinations conducted on these children between the ages of 7 and 11 years old.
PROFILER'S NOTE: Large proportions of the subjects were first children and had white
mothers; mothers and fathers were well-educated, with almost 50% having at least a four-
year college degree; and only 1 1% of the families had a yearly income below $20,000 at
baseline making limited applicability to the entire U.S. population.
Control group (n=227) included children from the cohort without fluorosis.
Birth to 36 months (and beyond)
Children were part of the original Iowa Fluoride Study and those included in this analysis
were those whom questionnaires were returned at ages 16, 24 and 36 months and had
received a mixed dentition examination. Fluoride intake was evaluated from water, food,
beverages, supplements, and dentifrices.
Fluoride intake of children from water, food, dietary fluoride supplements and fluoride
dentifrice was estimated from questionnaires completed by their parents at ages 16, 24 and
36 months. Fluoride intake (mg/kg B W/day) from water was estimated by multiplying each
subject's water intake by his or her water source fluoride concentrations and product-
specific water fluoride assay results. (See Analytical Methods below.) Parents reported daily
ingested quantities of water by itself, milk, ready-to-drink juices and juice drinks,
carbonated beverages, beverages mixed from powdered and frozen concentrates, foods
made with almost all water, foods made with some water and foods cooked in and absorbing
substantial amounts of water. The fluoride intake from beverages other than water and the
selected foods was determined by multiplying the daily intake in each category by the
average fluoride levels for those specific categories from the series of assays. Fluoride
intake from prescribed fluoride supplements was calculated using parental estimates of
frequency and dosage, paired with study data documenting the amount of fluoride contained
in the supplement. Estimates of the amount of fluoride dentifrice used were based on
parents' responses as to which diagrams of toothbrushes best depicted the amount that the
child used most often to clean his or her teeth and the estimated amount that the child
swallowed (<25, 50 or>75%). The fluoride intakes from beverages, selected foods and
dietary supplements were combined to estimate fluoride ingestion from sources other than
dentifrice. Medians, along with 25th and 75th percentiles, were calculated to summarize
fluoride ingestion (mg/kg BW/day) from dentifrice for the study sample. Cumulative
estimates (16 to 36 months) of fluoride ingestion from dentifrice and diet/supplements were
calculated using the area-under-the-curve (AUC) trapezoidal method (expressed as rng F/kg
BW per day).
Fluoride assays of water and most beverages were conducted using a fluoride ion-specific
electrode (Model 9609, Orion Research) and an ion analyzer (Model 920, Orion Research)
after using a total ionic strength adjustment buffer (TISAB II buffer, Orion Research) to
99
January, 2008
-------
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL
METHODS:
provide constant ionic strength, free up the fluoride and adjust the pH. Random samples
were read again at the end of the day to verify electrode accuracy, repeat readings on
standards to confirm the standard curve and analyze approximately 6 to 20% of samples in
duplicate. Mean reproducibility was 97 to 99%. Solids and selected beverages were
analyzed with a modified Taves hexamethyldisiloxane microdiffusion method, and the
fluoride concentration of the resulting solution then was assessed using the method
described above. More than 20% of samples were done in duplicate with mean
reproducibility of 98%.
As part of the Iowa Fluoride Study, children were followed from birth to 36 months with
questionnaires at 6 weeks and 3 months of age and then at three-, four- and six-month
intervals thereafter to estimate daily fluoride intake (mg/kg B W) from water, beverages,
selected foods, fluoride supplements and dentifrice. One mixed-dentition caries and
fluorosis examination was conducted on each subject between the ages of 7 and 1 1 years
(mean 9. 1 yrs) by one of two trained dentist examiners. This study used IPS data to describe
the influence of estimated fluoride dentifrice ingestion at ages 16, 24 and 36 months (both
individually and combined) on fluorosis experience in the early-erupting permanent
dentition. Total fluoride intake was estimated from parental questionnaires administered at
ages 16, 24 and 36 months. The reliability of the questionnaire was assessed for a portion of
the participants via follow-up telephone calls within two weeks of completing the
questionnaires. Telephone interviewers assessed recall agreements with regard to use of
fluoride supplements, toothbrushing frequency and type of water used (tap, bottled or both).
A total of 343 children (49% males and 5 1% females) were examined for fluorosis in
permanent teeth using the Fluorosis Risk Index (FRI; see NRC, 2006, page 90). The FRI
was adapted to include assessment of four enamel zones (the incisal edge/occlusal table, the
incisal third, the middle third and the gingival third) on each tooth, with the zones grouped
according to the age at which enamel formation was initiated.
Mothers provided information about their age, education, family income and number of
children in the household. Informed consent was obtained from the mothers before the
investigation; children provided assent before the dental examinations at ages 7 through 1 1
years.
Dental fluorosis on permanent teeth was determined at 7-10 years of age (mean 9. 1 years)
by one of two trained dentist examiners using the FRI. Before the examination, teeth were
dried. Fluorosis was differentiated from demarcated non-fluorosis opacities by the criteria of
Russell (1961). The FRI was adapted to include assessment of four enamel zones (the
incisal edge/occlusal table, the incisal third, the middle third and the gingival third) on each
tooth, with the zones grouped according to the age at which enamel formation was initiated
The gingival third zone was excluded from the analyses because it was not consistently
erupted. Fluorosis cases were defined as those in which two or more of the eight permanent
incisors had definitive fluorosis (according to FRI criteria in which a score of 2 is assigned if
50% or more of a zone has definitive fluorosis versus a score of 3 for severe fluorosis,
which also involves staining, pitting and/or other deformity). Non-fluorosis cases were
defined as those in which there was no definitive fluorosis in the permanent incisors. Dental
examiners conducted duplicate examinations in 39 subjects to assess interexaminer
reliability (using percentage agreement and kappa statistics), but the results were not
reported. Twenty-three subjects with one incisor with fluorosis were excluded from the
analyses.
Descriptive statistics were obtained for baseline measures and for tooth-brushing behaviors
at ages 16, 24 and 36 months. Medians, along with 25th and 75th percentiles, were calculated
to summarize fluoride ingestion from dentifrice for the study sample. Because fluoride
ingestion measures did not uniformly exhibit normal distribution qualities (the data were
skewed), tests of association used the Wilcoxon rank sum test (flurosis cases versus
nonfluorosis cases) and the Kruskal-Wallis test for multilevel baseline measures. Multiple
logistic regression models at each time point and for 16 to 36 months (AUC) tested
associations between fluorosis and fluoride ingested from diet and supplements. A joint test
was added to the combined intake for each model. P levels below 0.05 were considered
100
January, 2008
-------
RESULTS:
significant.
I
c
*L description of the study sample is shown in Table 1 directly from Franzman, 2006. Results
f the study in Tables 2 through 4 are shown directly from Franzman, 2006.
IHBLt 1
.T-V'^'-rX . ^™;~^~r " ^r+^yr.^W^t^;,;7^:?v«c:^1-*;>y
EXPUUMTMY
VARIABLE
SM
Mate
Inmate
Fire* Child?
V«
Ne
Moth*i i Bat*
White
Other
Annual Family IlKfliae*
t, $Z«,«M»
sa»,cn».siB,i»»
S3DflOO.SMi«S
S4«,ti«l-$4ft»»
$$0,W»'$5S,»i
iM8,«M»
mothei s E.linatiun'
High tthaoMGED* « less
S*nw wltef€
Two-year coll^§ degree
Foiir*y®^if TOM*9e 4&9»&
^raslnatefpr^fe^enaf school
father's E.liKstion'
Hlgls Mlwol^ED ®r leai
foni« &sW««|e
Twe-f «%r coll^f «r degree
Pmir-f ear SDlIsge stef r^e
HraslMat*/profeMloRaE ssliosl
SliBnli^r of PerftisiiMii Incfs^i's
With Elii.jrosis
None
On*
Two
Thn«
Pow
FVK to sis
S«¥*ts ts> sight
PERCEWTACE OF ALL
SUBJECTS
|M * S«*1
id
6t
42
««
98 6
l.S
11
14
22
20
a 3
t«>
i*
21
L&
29
20
26
17
11
as
20
Kii
a
IB
4
2
PERCENTAGE OF MOM-
FLUOROSIS CASES
(n a IJ.lt
•«
R;2
•12
E8
98
M
14
24
22
12
IS
16
21
Hi
27
20
^7
14
li>
2T
12
J0i>
11
34
1.4
34
17
21
20
£2
SO
is
0
t>
46
2S
l«J
»'*•
VALUE
i .00
.5*>
.66
.87
,.4C
4fl
™_J
^ TH^n'tV'thrw ef Ois 34S aubj*Bfc&' ^o b§d fiuorisift tai ccstj Jijsissjr cailv^srs Kwlud^l, fi'oni ilte aRsIys^s,
f AUlje-'clitkhbirIk
f OKI): ''^neral aej'Uiv^Iim*^ dijisma,
f: Not spji1»bk.
101
January, 2008
-------
TABLE 2
1
i
i
i
T
-
«
1
EX PL AMATORY VARIABLE
Ai HOIK j All Subfvcts
Brush teeth
AIIIOIMJ Subject* Wlu. Brush Teeth
Use fluoridated dentifrice
Use nonfluoridated dentifrice
Do no* use dentifrie*
Tftothbruching Fit-tjii^iicy (p*r Day)
Less than -etnce
Once
twice
Mere than twice-
AIIIOIKI Dentifrice tfeers
Us* dentifrice flavored for children
AtiHMKj Fluoridated dentifrice timers
Estimated amount *t fluoride per brushing
0,5§ mg* or more
Estirwated percentage of dentifrice
swallowed
SO
PERCENTAGE OF SUBJECTS BV A€£ fW * 343)
1C MONTHS
90
es
32
as
1.4
i
45
11.
IB
S2.
24 nWIUTHS
too
2
8
25
Si
23
SO
29
7"
9
85
X MONTHS
too
86
i.
3
,18
fe'7
24
J.
60
45
21
1:3
66
'* Source: Franzman and otilloagues.*
t mf : MilHgraj»*.
PABLE 3
ftJM%
ICi AT
1ENTIFP.KE
MGESTntt
morjTHs}
6
4
K
S-B© JiSIC'
,\. i.'i.-:?',,. -..^-,..=.... .,, ^..,-..^v,. s v, ,,;il, Ni..^..^^.':''^.!.^^.,..
'i.f ! ! i--*3 * j '!? t.-''; if t'j •;.'! '''si !&1 f't is -I'-i i!r ,f.6
22tt OJH1 fO.Onft, O.O20.'
VH(TM«>«.imh.»l
No. -of Median dally BSth «*h_
S» O.002 (0.000,0.010- .61
««• 0.01,7 iorwe, o.o3S> 02
80 o.nie roooT, o.osb .02
86 0.0 IS (O.009, 0.038 1 .OS
Baaed iaa Wilccocon rsfik .cum tsst.
Weight dala w&re ijussing &' f-Mjt" of 3S sulgedx
AUC: ATM under the mm.
102
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE
NOT FOUND IN NRC
(2006)
PROFILER'S Initials
REMARKS /Date
VAD/12-
30-06
PROFILER'S ESTIM.
NOAEL
PROFILER'S ESTIM.
LOAEL
SUITABILITY FOR DOSE
RESPONSE MODELING
CRITICAL EFFECTS:
TABLE 4
%:. B^'&>.^t: "iJi-ivfo KT;l±j-£v!::
MOBIL FLIJORIDE-
M1URCE
16 Monlhi 1M.--1 and «](iph #F 1 sunKhntl tkn-iatiiua
5 A'U* ': Ar&rf uncte tla& curve.
^:^-mm^
OF> ADJUSTED
ODBS RATIO!
1 o!n7
i i m
-1 I .-?€»
i i «*>;
i i,s3
i i *<
.:^:^^^;uikk;;;^tj-c,:.:_: _
95 PERCENT p VALUE FOR i» VALUE MR
CONFIDENCE COMPONENT COMBINED
INTERVAL INTAKE
O'.-.I-I'K !*! lfl
(J.9T-J f-9 ,10 nn.7
l.iKi-t.ii*. M ()07
1 lM-2',1." ,0001 ))WI(
i'wM'fil •*o? •jf QUMM* ss^wmk-d wth a
The study authors concluded that the results of the study affirm the link between use of
fluoride dentifrice and the development of mild dental fluorosis; in particular, the data
suggest that this relationship is most pronounced at about age 24 months. The findings
suggest that health professionals need to emphasize the proper use of small quantities of
dentifrice among toddlers, with appropriate parental supervision, to limit the amount of
ingested dentifrice.
Hong, L., S.M. Levy, J.J. Warren, B. Broffitt, and J. Cavanaugh. 2006a. Fluoride intake
levels in relation to fluorosis development in permanent maxillary central incisors
and first molars. Caries Research. 40:494-500.
Hong, L., S.M. Levy, B. Broffitt, J.J. Warren, M. J. Kanellis, J.S. Wefel, and D.V. Dawson.
2006b. Timing of fluoride intake in relation to development of fluorosis on
maxillary central incisors. Community Dent Oral Epidemiol 34:299-309.
Russell, A.I. 1961. The differential diagnosis of fluoride and non-fluoride enamel opacities.
J Public Health Dent 21 : 143-146.
Large proportions of the cohort were first children and had white mothers; mothers and
fathers were well-educated, with almost 50% having at least a four-year college degree; and
only 1 1% of the families had a yearly income below $20,000 at baseline. Therefore, results
are not representative of the general US population. Parents reported the data and
investigators did not directly observe the toothbrushing behaviours of children. Some
information was missing as a result of some parents' failure to return questionnaires, which
may have affected the results. Severe fluorosis was observed in only 3 of 39 subjects;
therefore, it was not possible to assess the association of dentifrice use with this more
involved dental fluorosis.
The study design did not identify
a no-fluorosis intake dose.
The study design did not identify
a lowest fluorosis intake dose.
Not suitable,®; Poor (J; Medium (_); Strong (_)
Not suitable for dose response modelling as the amount of fluoride intake associated with
the three cases of severe fluorosis (which is considered to be adverse) was not given. The
study could be used, however as a relative source contribution as to the effect that fluoride
from dentifrice ingested at a young age can affect the incidence of fluorosis in permanent
dentition.
Dental fluorosis (permanent incisors)
103
January, 2008
-------
Galagan, D.J. and G.G. Lamson. 1953. Climate and endemic dental fluorosis. Public Health
Reports. Vol. 68, No. 5: 497-508.
ENDPOINT STUDIED:
Dental fluorosis
TYPE OF STUDY:
Cross-sectional survey
POPULATION STUDIED:
US/Arizona; Fourth through ninth grade (ages 9-16) children from public and parochial
schools in 6 Arizona communities made up the study population with 83% being of
Spanish descent. To be included in the study, the children were required to have exposure
to the water in that community from birth through age 9. The following table copied
directly from Galagan and Lamson (1953) provides the number of children in each city
and the age during the study. No distinction was made between male and female children.
Number of children in
each community
Yuma {82}
Tempe (113)
Tucson (316)
Chandler (§5)
Casa Grande (50)
Florence (70)
A
9 •
5
5
3
8
8
3
ge (years) and percent In
each age group
10
9
18
7
15
12
9
11
7
16
-8
14
22
9
12
10
18
11
20
12
16
13
7
il
18
15
18
18
14
18
12
24
9
14
14
15
16
13
17
12
8
20
16
28
12
12
7
6
11
CONTROL POPULATION:
None
EXPOSURE PERIOD:
A requirement for children to be included in the study was exposure to the community
drinking water from birth through their 9th year.
EXPOSURE GROUPS:
The following table was constructed by combining data in two separate tables in Galagan
and Lamson (1953) to indicate the fluoride levels identified in the water, the source of the
water supply and the number of samples taken from each community.
Data for children in Arizona communities fluoride study
Community
Yuma
Tempe
Tucson
Chandler
Casa Grande
Florence
Source of
supply
Colorado
River
4 wells
17 wells
4 wells
5 wells
4 wells
Treatment done
Desilting, Al SO4, flocculation,
CuSO4, filtration, chlorination
chlorination
chlorination, ammoniation
None
Do.
Do.
#of
samples
79
7
31
16
20
22
Mean
fluoride
content
0.4
0.5
0.7
0.8
1.0
1.2
PROFILER'S NOTE: Although not stated in the table, the profiler assumes the mean
fluoride in the table should be in "ppm".
Some explanations about disruptions to water supplies and/or changing wells were
explained in the study article and are included below.
Yuma: In the summer of 1937, sewage backed up in the Colorado River making the
drinking water unsafe. Water for drinking was redirected from a well that was used to fill
the city's municipal pool for 3 months. Fluoride levels from the pool well were not
104
January, 2008
-------
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
obtained during the time of its use, however, a sample in May 12, 1951 indicated a
fluoride concentration of 0.6 ppm. This indicates that possibly for a 3 -month time period,
the 16-year old children would have been exposed to 0.6 ppm instead of 0.4 ppm.
Tucson: The city of Tucson has two different systems of drinking water; one for the
Northside and one for the Southside. Because the Southside has a higher amount of
fluoride, the study used children well within the boundries of the Southside system. The
two systems have had two sets of controls since 1938, so priorto 1938, there might have
been some cross- mixing between the two systems. Since the Northside fluoride is 0.3
ppm, the result would have been exposure to a lesser amount of fluoride for a short time.
Chandler: Two of Chandler's four wells had their casings break in 1938 and water from
outside entered the wells. Replacement wells were added, one in 1944 and one in 1948.
The mean of fluoride concentrations taken with the original wells was 0.75 ppm and with
the new wells, 0.85 ppm. Fluoride levels were only checked once during the casing
breakage when outside water was entering the wells and it was 0.8 ppm. These values all
indicate a fairly stable fluoride concentration in the area.
Case Grande: The five wells in this community were put into place in 1922, 1930, 1937,
1946 and 1950. The addition of new wells did not appear to affect any of the fluoride
levels as twenty analysis from 193 1 to 195 1 have ranged from 0.9 to 1.2 ppm.
Florence: During the study period, the community used 4 wells; two were originally from
1919. A new well replaced these two in 1939 and another well was added in 1947. A total
of 2 1 fluoride samples were taken with 14 taken prior to July 1 93 9 and 7 taken after July
1939. Both averaged 1.2 ppm of fluoride indicating the well change did not affect the
fluoride levels.
PROFILER'S NOTE: Although there have been well replacements and changes to some
systems, overall, the fluoride levels were not adversely affected, and no confounding
evidence was identified.
Only fluoride exposure from drinking water was considered in this study. However,
Galagan and Lamson (1953) did state that the study children lived in desert cities with a
mean annual temperature of 70°F, a mean relative humidity of 37%, 80-85% sunshine and
extremely high day-time temperatures which all could have increased the amount of water
the study population drank. The authors did note that field irrigation is common in Arizona
thus possibly allowing fluoride to be absorbed into the crops and that most of the children
in Arizona were of Spanish descent with beans making up a large portion of their diet.
Most of the beans, in cooking preparation, are soaked in or boiled in water for long periods
of time and could create another source of exposure. While both of these were noted, the
authors did not think the soil contamination was substantial and the scope of the paper did
not include the dietary exposure possibilities.
Fluoride levels for the water were analyzed intermittently from 1935 to 1950 and quarterly
through 1951, although the method for analysis used was not provided in the study. No
other chemical parameters in the water were provided.
Children were chosen from the 4th to 9th grade in private and parochial schools within the
study communities. Children were only included if they had been residents of their
community. A questionnaire confirming the residency was sent home to the children's
parents and/or guardians. If residency was confirmed, the children were scheduled for a
dental examination. All children were examined by the same examiner. Examinations
were performed with the children seated in a dental chair using a mouth mirror and Burton
light. Compressed air was available, if needed. Each tooth was given a fluorosis score
based on Dean's Index (1942). These classifications were later put into a child-
classification and used to compute the community fluorosis index (CFI).
Dean's Index (1942) was used to rate fluorosis on the teeth of the study children. The
105
January, 2008
-------
MONITORED:
community fluorosis index (CFI) was also calculated according to Dean (1942).
PROFILER'S NOTE: Dean's Index (1942) is described in Section 2 of the report.
STATISTICAL METHODS:
No statistical methods were included in the study report.
RESULTS:
Dental fluorosis
Table 4 was copied directly from Galagan and Lamson (1953) and indicates the
prevalence and severity of fluorosis in the children within each Arizona community. The
results show that as the fluoride concentration rose, the community fluorosis index
increased and the number of children without visible fluorosis decreased.
Table 4, Prevalence of fliwresis, distribution of signs of fluorosis an«l community fluorosis indexes
in rotation to fluoride concentrations of common water suppltes continuously used by 726
children examined in six Arizona communities, 1951
Community
-
"Hieson.. _„
Tliandler,,,,^
Fluo-
ride
con-
sen-
tra-
tion
0 4
5
7
8
i n
Florence-
nber
lit-
ren
am-
ed
82
113
816
9fi
60
70
Number
chil-
dren
affected
3
It
53
18
24
39
Xumber of examined children witli signs of fluorosis
Fluorosis absent
'Norfflal
63
59
120
40
7
17
Ques-
tion-
able
26
43
143
37
19
14
Fluorosis present
Very
mild
2
10
38
9
15
18
Mild
10
6
9
10
Modejv
ate
a
9
Severe
1
2
Com-
munity
fluo-
rosis
index
0.21
.30
.48
.52
.85
1. 12
PROFILER'S NOTE: The children in Casa Grande (1.0 ppm) did not show the same trend
toward increasing severity in fluorosis that the other communities indicated, although the
CFI did. This is possibly due to the smaller number of children that were examined in the
area as compared to the other communities.
Comparisons of effects in
areas with different
climatological variables
Table 5 was copied directly from Galagan and Lamson (1953) and indicates the
prevalence and severity of fluorosis in the children within each Arizona community as
compared to ten Mid-west communities (Dean, 1946) with similar fluoride concentrations
but having very different climatological variables. Figure 3 was also copied directly from
Galagan and Lamson (1953) and shows the relationship between the fluorosis indices of
the different communities. The Mid-west communities used in comparison have a mean
average annual temperature of 50°F, whereas, the Arizona communities have one of 70°F.
The chart shows that the Arizona communities had a higher percentage of children
affected by fluorosis when compared to cities with similar fluoride levels but cooler
temperatures. The figure shows that the communities in Arizona have a much steeper
slope and the index line accelerates at twice the rate of the Mid-west communities into the
borderline zone (CFI value of 0.4- 0.6 ppm). Dean stated that a CFI of below 0.4 ppm has
little or no public health concern, 0.4-0.6 ppm is borderline and >0.6 ppm means that
excess fluoride should be removed from the water. The figure suggests that the children in
Arizona would develop twice as much fluorosis as the Mid-west children, although
discrepancies between the two examiners must be considered. The author concludes that
the higher temperatures in the Southwest and the increased amount of sunshine present per
day (radiant heat) must help contribute to the increased fluorosis present by causing an
increase in water intake.
106
January, 2008
-------
Table 5. Prevalence of fluorosis percentage distribution of signs of fluotesis ond community
floorosis indexes in relation fo fluoride eoncerrtralion of common -water supplies of 16 torn-
„ . reunifies in two temperature zones
Percentage distribution of signs of flwc
Marion, Ohio—
Elgin, 111
Pueblo, Colo
Kewanee, 111.
Aurora. !!!„„ ^-\
East Moline, lit I
May-wood, IB
Joliettlll
Elmhurat, In.
1 A-veraet- Ariinra B0.3" P.; Middle West 50,«° F.
s keas than 0,5 peu'eiit.
NOTE: Age matt far 4m»na group, 8-16 yeurt; midwMtein groi,p
[Jrid'd
n: Flnnrifts tltU im
t fnr Die A
,
1* j«»rs
cnnuniiiiities fmrn Pesn, H, T..
es st.d lluorue
irt'm j 'tm>u 11* s
IJrUi'd rjt»tt«, Anwri'ar AsM'cUfiurt tnr ine .wvaiwcntenc of Sci«.t*: Bfit'Rl >.o^>« i..... ..""•••" „.',"";•"
LsufMtar. Pa., Wfi; IT.MI atmi-a) k-»>rftat.m fr,C ArtOTnft w-wmunitiw from, ME.tb il \ The [). Bn ipTUi N«. 19,, Julv i94,H; for ro.dv,
c«.«I fr™ T. S Oepartinert ut t'ommprf?, Wither Bm-a«: Uouatc-foraw] data, SfunlWy atirl a
«'imma-ics- f hi- im-ar. am-iual temperature fur Aurnr* Juliet, «t.d Fl/tiu is repre«>js1rd by tl,2 13 year >rt«r»ge
tupiperatun- mt A n-ra; tlmt f'-r Ken Hi.rw, Eakt Miiiit.t, ard {5»]««.bi,ru liv ti* IP-J'Pflf avernRe me-»n itmpc
fur O*lv* and that 'IT Msmn-ood am* Uli-ihnr'.t t.y ti.u IU-yt»r iveregt nie»n n-mpErrtniefur ' >•'—"•
r? 3. R^E^fiQ£ishi)3 b^iw€€ra Hu0rid^ concenlrotion of munlci|^cil waters and Hn@fO$!$ intl&x far
communities with mean annual temperatures of approximately 50° F. (Mid westl and 70' F.
(ArJzenn)
2.0
PROFILER'S NOTE: The trend toward higher fluorosis was observed in the Arizona
communities at a lower fluoride concentration; but there were no details provided to
indicate whether the fluorosis occurred in primary or permanent teeth. The data from
Arizona included children as young as 9 indicating some still would have primary teeth;
the data from the Mid-west was for children 12-16 years old indicating mostly permanent
dentition. To get a true picture of comparison, only the older Arizona children should have
been compared to the older Mid-west children. The study does not indicate whether the
children in the Mid-west had the same time of exposure (birth through 9 years) as the
Arizona children.
107
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S DFG/2-13-07
REMARKS
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
Galagan and Lamson (1953) summarized that in the Arizona communities studied,
fluoride concentrations above 0.8 ppm resulted in objectionable dental fluorosis, 0.6 to 0.8
ppm resulted in occasional diagnosis of fluorosis and concentrations below 0.6 ppm
caused no objectionable dental fluorosis. Also, when comparing the CFFs, fluorosis
occurred at about twice the intensity in Arizona communities when compared to Mid-west
communities with comparable fluoride concentrations but different climatic factors. As a
result of the increased temperature and radiant heat gain, Arizona children drink more
water than those living in more temperate climates causing increased ingestion of fluoride
in relation to the concentration found in the water supply.
Dean, H.T. 1946. Epidemiology studies in the United States. In Dental caries and fluorine,
edited by F.R. Moulton. Lancaster, Science Press, p. 5-31.
The survey clearly shows a difference in the amount of fluorosis identified based on
climate changes, although a stronger study would involve the same examiner observing
children from each area to prevent examiner differences and comparisons of the same aged
children. The profiler agrees that different acceptable fluoride levels might be needed
based on the area of the country one inhabits. An examination into the food fluoride
content should be conducted in this community to include some of the food preparation
habits that the children were exposed to. Also, a true measure of the drinking habits of the
children in the area would also help support the finding of increased exposure to ensure the
children actually do drink more water than the Mid-west children. The study lacked any
statistical analysis.
Although the study is limited by a small number of data points, a NOAEL of 0.7 ppm was
established based on no children exhibiting severe fluorosis at this concentration.
Although the study is limited by a small number of data points, a LOAEL of 0.8 ppm was
established based on the number of children with severe fluorosis.
Not suitable ( ), Poor (), Medium (X), Strong ( )
While the study is not ideal due to the small number of data points and the lack of data on
true water intake, it does indicate an increase in fluorosis severity as fluoride
concentrations increase and could be used in conjunction with other studies to provide
weight-of-evidence to set some guidelines in areas with similar climates. One could also
use the CFI values from the study to set up guidance for other similar geographic areas.
Dental fluorosis based on climate variables
108
January, 2008
-------
Gopalakrishnan et al., 1999. Prevalence of dental fluorosis and associated risk factors in Alappuzha
district, Kerala. Nat. Med. Jour. India, Vol. 12, No. 3: 99-103.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
Dental fluorosis
Community-based, cross-sectional study of fluorosis and fluoride levels in drinking water.
India, 1142 school children (630 girls, 512 boys), ages 10-17 yrs.
None
10-17 yrs, from birth to time of oral examination of teeth in January & February, 1998
Drinking-water fluoride levels based on data from the water department of Alappuzha. The
exposure groups were 0.7, 1.1, 1.2, and 1.4 ppm.
Drinking water was the main exposure route evaluated. However, brick-tea intake, sea fish
intake, dry-fish consumption, and toothpaste used for dental cleaning were also evaluated.
Not stated what analytical method the water department used to analyze for the fluoride
levels in the drinking water.
The prevalence of dental fluorosis was studied in Indian school children, ages 10-17 yr, and
the contribution of potential risk factors was evaluated. A dental specialist evaluated each
child for the presence and severity of dental fluorosis, using a modified Dean's index of
Fluorosis (see below), as well as any other dental conditions. Printed questionnaires were
given to the students 1-2 days before the dental examinations (to be returned at the
examination) with questions concerning socioeconomic status, occupation and level of
education of the parents, the source of drinking water, amount of water consumed, brick-
tea consumption, sea-fish intake, and use of fluoride-containing toothpaste by the student.
The fluoride content of the water was obtained from the local water department. The
prevalence of dental fluorosis was estimated by taking all cases of dental fluorosis as the
numerator and the total child population evaluated in the age group of 10-17 years old as
the denominator. A Community Fluorosis Index, as described by Dean (1942), was
computed.
The study authors reported that Dean's was used to evaluate the grade of dental fluorosis;
however, as described, the scoring system was modified to include only five categories:
normal, questionable, mild, moderate, and severe, and furthermore, the severe level was
described as cases where white areas covered more than 50% of the surface areas of the
teeth; no mention is made of pitting or brown staining.. A Community Fluorosis Index, as
described by Dean (1942), was computed.
The association of dental fluorosis with select individual risk factors was studied using Chi-
square and Chi-square trend tests. Multivariable logistic regression (with step-wise forward
selection) was used to evaluate the independent association of select risk factors with the
prevalence of dental fluorosis. A p value of <0.01 was used for entry of the variables in the
multivariable models. Odds ratios (and their 95% CI) for the association of the predictor
variables with the dependent variable were computed. A p value of O.05 was taken to
indicate statistical significance.
See Table II for the prevalence of dental fluorosis in the study sample, Table
III for the prevalence of dental fluorosis according to the source of drinking
water, and Table IV for the variation in prevalence of dental fluorosis in the
different regions according to the fluoride content of the drinking water.
109
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/date
SBG 3/27/07
TABLED. Prevalence of denial fluorosif « »udy sample
VButun in fiMvitaM No rf Piwiteaw of dtoul flmom {%) 3 «h»
ebii4«B Ko«m»l Ondel Graded tlndrin Ondt IV AUGndetl*)
TMilHupk »»« "»» '«.4> ISI (U,l) III (18.M ?i 'i.2| if ISJ) IS.*
^rwrrfwf to |?l0f« **/ wid^w*
Urha JftO :$' <**« 180 07 M il«H>» <' ilOl) 3S3 <8f»l«
Rani $$; 4«4 (»2> 51 (SI) 27 (46! '0 (! ?! 10 if. 168
Actw*fet£ W» g*i*der
jjoyt MI W (M«l 62 (121) ** (9*1 K I" *) i* I'D H 3 «fl(M"
Girh « JM (MS) W tM.ll 70 till) « s 2 H T) 326
* Ort s^«wt leal t Chi a|isMi ^»*5^ s*«
T*»lt III Prcviknce of denul fluncosri accocding to the souree of douking witet
S«st« isf dtnukMi Ns, 0f tevjileMS sf efenui ^yo»&n t^l* t p vtl«
««Ercon«ioKi tblldnn Noraiil (*«* 1 Grade II C3f«fc W Ondf IV AllOndetm
TutawellvMn 161 (»»)! 3 Oi 4 (4) 2 0 <0) 3 (13.«i 364
Only pipe MUT «06 44? (11.J) ISO (IA n 1 05 03) « (8.H 61 (?,«» 44 i
* CSi-sqasit wofl wi
TABU; IV, Vanation in prevalence of tkiftnl fluorosis in panchavats according KS thg fluoride cJ4i 0 lO) I (161 9 il*» (1 («; J«
Ms^Mficfecn
ft»«tttd 87 »' ''It U1') » C"! I 112! J 14 1) 1 <*!> (61 «OM«*
AnA^ppirlu ; 1 IK Ml i*M» « itS»» 1 |«>4t » (62' I «2i HB
fp»«fj» 1J J>1 l«< «l»l) Z» 1103) "> <»7t > (63) 7 -34i 10.4
Aiym* 1* 5k 4< .»!«) 5 >47) 1 iIA 3! 2 |7l 0 ifl] 241
" t,^» «gt*ra« nsiwf lass
The authors noted a significant positive association between water fluoride content and
prevalence of dental fluorosis. A step-wise increase in the prevalence of dental fluorosis
was noted with a corresponding increase in the water fluoride content in different regions.
The socioeconomic status of the parents, brick-tea drinking, sea- or dry-fish consumption,
and the use of toothpaste were not associated with the prevalence of dental fluorosis.
None
This study was in India and there was no control group with low fluoride levels in the
drinking water, so it would not be representative of the U.S. population. In addition, the
authors assumed that the fluoride content in the water had not changed over the last 15
years (when fluorosis occurs during early childhood during the time of calcification of the
teeth), which could be a big assumption.
The fact that severe fluorosis occurred in the study population exposed to drinking water
fluoride concentrations of 0.7 to 1.2 ppm, but not at 1.4 ppm suggests that other factors
were contributing to total fluoride exposure. The authors state, however, that brick tea
consumption, ingestion offish and the use of toothpaste were not positively associated with
fluorosis (quantitative data were not presented in the publication). The inconsistency of the
results reduces the value of the data for dose-response modelling, but it does provide
evidence that severe fluorosis can occur at low water fluoride levels in hot tropical
climates.
110
January, 2008
-------
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
Based on the data presented in Table IV, there does not appear to be a NOAEL for severe
dental fluorosis (Dean's Index of 4) identified in this study.
Based on the data presented in Table IV, the LOAEL for severe dental fluorosis (Dean's
Index of 4) appears to be 0.7 ppm.
Not suitable (_ ), Poor (x ), Medium (_ ), Strong (_ )
This study seems to be poorly suited for dose-response modelling because of the
inconsistency of the results, which may have been due to the fact that the range of fluoride
concentrations was small (i.e., 0.7 to 1.4 ppm). Furthermore, for the level of severe
fluorosis there was no clear dose response.
Dental fluorosis
111
January, 2008
-------
Grobler, S.R., Louw, A.J., and Van W. Kotze, T.J. 2001. Dental fluorosis and caries experience
in relation to three different drinking water fluoride levels in South Africa. International Journal
of Paediatric Dentistry. 11: 372-379.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
Dental fluorosis; caries
Cohort study
South Africa/ Leeu Gamka: 120 children (45 girls, 75 boys), aged 10-15 years old, and
lifelong residents. The average altitude and annual rainfall are 550 m and 150 mm,
respectively.
South Africa/Kuboes: 115 children (63 girls, 52 boys), aged 10-15 years old, and
lifelong residents. The average altitude and annual rainfall are 200 m and 100 mm,
respectively.
South Africa/Sanddrif: 47 children (3 1 girls, 16 boys), aged 10-15 years old, and
lifelong residents. The average altitude and annual rainfall are 22 m and 50 mm,
respectively.
The children in all groups had similar ethnic and socio-economic status, nutrition and
dietary habits, and virtually no dental care or any fluoride therapy. The subjects were of
mixed ethnic origin, originating from Khoi, Caucasian, and Negroid roots, developed
into a homogenous ethic group over hundreds of years. The socio-economic status was
low, as reflected in residence in subeconomic housing units. The staple diet consisted
of bread and potatoes, with occasional intake of other vegetables and meat. No dietary
habits were flagged that would significantly contribute to fluoride ingestion, including
the use of naturally occurring salt deposits containing high levels of fluoride.
The study group in Sanddrif served as the control population.
From birth to age 10-15 yr. The dates when the examinations were conducted were not
reported.
Leeu Gamka water supply (boreholes) naturally fluoridated at 3.0 ppm (range 2.7-3.3
ppm).
Kuboes water supply (boreholes) naturally fluoridated at 0.48 ppm (range 0.45-0.50
ppm).
Sanddrif water supply (Orange River) naturally fluoridated at 0.19 ppm (range 0.15-
0.23 ppm).
The fluoride levels correspond to averages "over the last ten years", with at least one
sample analyzed each year. NOTE: children were 0-5 yr old when the first
measurements were taken.
The study authors state that communication with personal health care personnel
indicated that there were no dietary habits that may have contributed to a significant
exposure to fluoride, and it was reported that the children in this region had virtually no
dental care and or fluoride therapy, including the use of fluoride dentifrices.
The water fluoride level was determined potentiometrically according to the method
described by Nicholson and Duff (1981) and was analyzed over a period of
approximately 10 years with at least one sample per year.
112
January, 2008
-------
STUDY DESIGN
The study included 282 children aged 10-15 years with virtually no dental care or any
fluoride therapy who lived continuously since birth in three different naturally
fluoridated areas of South Africa: Leeu Gamka, 3.0 ppm F (n=120); Kuboes, 0.48 ppm
F (n=l 15); and Sanddrif, 0.19 ppm F (n=47). The children in all groups were of similar
ethnic origin and low socio-economic status; nutrition and dietary habits were similar.
Dental examinations were made by the second author using portable dental equipment.
None of the children refused to be examined. The teeth were examined for caries and
fluorosis using DMFT and Dean's indices, respectively, according to WHO guidelines.
Caries criteria of the WHO (detectably softened floor, undermined enamel, and/or
softened wall) was strictly adhered to in order to distinguish between hypoplastic
fluoride lesions and carious cavities; where any doubt existed, caries was not recorded
as present.
Examinations: Examination details were not reported, including location where
examinations occurred (assumed to occur at school), lighting conditions, or equipment
used.
The examiner was standardized and calibrated for intra- and inter-examiner variability
prior to and during examinations; agreement was determined using weighted kappa (k).
The intra- and inter-examiner agreement scores for DMFT (k=0.90 and 0.85) and
fluorosis index (k=0.78 and 0.78), respectively, were substantial to almost perfect,
according to the scale of Landis and Koch, meeting the scientific requirement for
validity and reliability. Agreement was also monitored by re-examining 10% of the
sample, with the same result as was found in the pre-survey calibration finding.
PARAMETERS
MONITORED:
Dental caries and fluorosis were measured using the mean number of decayed, missing,
and filled permanent teeth (DMFT) and Dean's indices, respectively, according to
WHO guidelines. Fluorosis was scaled as none (0), questionable (1), very mild (2),
mild (3), moderate (4), or severe (5). No radiographs were taken during the surveys.
STATISTICAL METHODS:
The Mann-Whitney [/test was used to analyze the data for differences between males
and females. The Chi-square test was used to detect differences in age distribution in
the fluoride areas. DMFT scores were analyzed by the Kruskall-Wallis test and the
Bonferroni test. Correlation between DMFT scores and fluorosis was analyzed by the
Spearman correlation test. Significance was set at p <0.05.
RESULTS:
Caries
Tables 1 and 2 were copied directly from Grobler et al. (2001). Table 1 summarizes the
mean age (years), DMFT, percent caries-free, mean fluorosis score, and percent
fluorosis-free children in each community. No significant difference was found
between males and females for each parameter tested (Mann-Whitney U test), so the
results for the two series were combined. The mean age in the three areas did not differ
significantly, although a difference in age distribution was found (Chi-square test), with
younger children in the Leeu Gamka area.
The mean DMFT scores for the children in Sanddrif and Kuboes were the same
(1.64±0.30 and 1.54±0.24, respectively), but significantly higher (1.98±0.22) for Leeu
Gamka (Kruskall-Wallis, Bonferroni test). The percentage of children that were caries-
free was 47% in Sanddrif, 50% in Kuboes, and only 29% in Leeu Gamka, the high
fluoride area. The proportion of caries-free children in Sanddrif and Kuboes did not
differ significantly and was significantly higher than Leeu Gamka. The decayed
component dominated the DFMT score in all three areas, with significantly (Kruskall-
Wallis test, p<0.01) more children affected in the high fluoride area (Leeu Gamka)
compared to the other two areas. Both the filled and missing components for all three
areas were almost non-existent and did not differ significantly. A strong positive
correlation (Spearman correlation test) was found between caries experience and the
fluorosis scores of children in the high fluoride area, but there was no correlation in the
other two areas.
113
January, 2008
-------
Dental fluorosis
STUDY AUTHORS'
CONCLUSIONS:
Table 1, Nfcaa age 2) was 47% in Sanddrif, 50% in Kuboes, and 95% in Leeu Gamka. The
proportion of children with fluorosis (scores >2) did not differ significantly between
Sanddrif and Kuboes, but was significantly lower than Leeu Gamka. Approximately
half the children in both of the low F areas had no (or questionable) fluorosis (scores
<1) compared to only 5% in Leeu Gamka. The prevalence of moderate to severe
fluorosis (score >4) was 4% and 6% in Sanddrif and Kuboes, respectively, but was 61%
in Leeu Gamka, the high F area.
T«fcle 2. Dental iuorosis i*ore by fluoride »rea.
Sioddrif Kuboes Lceu Giwka
1>nM| (F-0'Wp.p.m.) (f- 0-4-8 p-p.m,> > (p.p.ro.)
fluorosis score n % a % a %
0 - Normal !8 38-3 46 40-0 I 0-8
1 - Questionable 7 11-9 11 9-S 5 4-1
2»V«ymiM 12 JS-5 39 33-9 W 15-8
3 - MiM 8 11-0 12 10-4 22 1J-J
4 = Moderate 2 4-2 6 5-2 J7 JO-8
S - Severe 0 0-0 1 « 36 JO-0
Total 4? 115 120
PROFILER'S NOTE: The profiler agrees that no difference in fluorosis was found in
the two low fluoride areas (Sanddrif and Kuboes), but the fluorosis score was
comparatively higher in the high fluoride area (Leeu Gamka). This trend was also noted
in the percentage of fluorosis-free children; no difference between Sanddrif and Kuboes,
but comparatively lower percentage fluorosis-free in Leeu Gamka.
The percentage of caries-free children for both lower fluoride areas was higher, with
fewer children caries-free in the high fluoride area (Leeu Gamka). Both the fluorosis
and DMFT values for Sanddrif and Kuboes did not differ at all. The breakdown in
fluorosis scores (Table 2) and DMFT (Table 1) for Sanddrif and Kuboes showed that
the difference in the drinking water supplies for these two areas did not make any
significant difference whether it was 0. 19 or 0.48 ppm fluoride. Most (61%) of the
children from the high fluoride area (3 .0 ppm F) were classified as having moderate to
severe dental fluorosis in comparison with only 4.2% and 6% for the areas with 0. 19
ppm F and 0.48 ppm F in the drinking water, respectively. Teeth with moderate and
severe fluorosis had dental caries more frequently than teeth with no or very mild to
mild fluorosis.
114
January, 2008
-------
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/date
SJG/3/19/07
PROFILER'S ESTIM. NOEL/
NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
The results suggest a positive association between high fluoride levels in the drinking
water and dental caries. Furthermore, a low caries experience and no difference in
DMFT and fluorosis between the two low fluoride areas were found.
Landis, J.R. and Koch, G.G. (1977). The measurement of observer agreement on
categorical data. Biometrics 33: 158-174.
Nicholson, K. and Duff, EJ. (1981). Fluoride determination in water. Analytical
Letters 14A7: 493-5 17.
World Health Organization (1987). Oral Health Surveys: Basic Methods. 3rd ed.
WHO, Geneva.
Overall, the study was well-conducted and had adequate study design. However, it
should be noted that the sample size was not equivalent in all areas, which might skew
the statistics if not considered; the area with the lowest fluoride level in the water
(Sanddrif) had a lower sample size (n=47) compared to the other areas (n=l 15 and 120).
The prevalence of caries was comparatively higher in the high fluoride area (Leeu
Gamka) compared to the other two areas (Sanddrif and Kuboes). A lower percentage of
caries-free children was noted in Leeu Gamka compared to Sanddrif and Kuboes. The
fluorosis score was higher in the high fluoride area (Leeu Gamka) compared with the
two low fluoride areas (Sanddrif and Kuboes). A lower percentage of fluorosis-free
children was found in Leeu Gamka compared to the other two areas. The prevalence of
fluorosis was much higher (95%) in the high fluoride area compared to the other two
areas (47% in Sanddrif, 50% in Kuboes); this included a higher prevalence of moderate
to severe fluorosis in Leeu Gamka (6 1%) compared to Sanddrif (4.2%) and Kuboes
(6%). No significant differences were found between Sanddrif and Kuboes with respect
to caries experience or fluorosis prevalence/severity. A higher DMFT score correlated
with a higher mean fluorosis score.
Factors, noted by the study authors, that could influence the findings include:
1) Temperature; high average maximum daily temperatures (~25°C) result in
elevated water consumption.
2) Age; caries experience increases with age. Leeu Gamka had younger children
compared with Sanddrif and Kuboes, negating the possibility that age (as
opposed to the high fluoride content in the drinking water) might have
contributed to the higher DMFT found in Leeu Gamka.
The information presented is useful for evaluating the occurrence of severe fluorosis in
populations whose fluoride exposure was not affected by dietary habits or the use of
fluoride supplements or dentifrices. Confounding factors are the low socio-economic
level of the population, the possible effects of poor dental hygiene, and the arid climate
which is likely to increase water consumption and overall fluoride intake. Nevertheless,
the data may be useful for comparison with US populations studied in hot arid climates.
Study design was not suitable for development of a NOAEL for caries experience or
threshold dental fluorosis. The NOAEL for severe fluorosis was 0.19 ppm
Study design was not suitable for development of a LOAEL for caries experience or
threshold dental fluorosis. The LOAEL for severe fluorosis was 0.48 ppm (0.8%
incidence).
Not suitable (_), Poor ( ), Medium (X), Strong ( )
The study indicated a clear dose-response for the incidence of severe fluorosis (0% at
0. 19 ppm, 0.08% at 0.48 ppm and 30% at 3.0 ppm.
115
January, 2008
-------
1 CRITICAL EFFECT(S):
Prevalence and severity of dental fluorosis and caries experience |
1
116
January, 2008
-------
Heifetz, S.B., W.S. Driscoll, H.S. Horowitz, and A. Kingman. 1988. Prevalence of dental caries and
dental fluorosis in areas with optimal and above optimal water-fluoride concentrations: a 5
year follow-up survey. J. Amer. Dent. Assoc. 116:490-495.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL
POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL
METHODS:
STUDY DESIGN
Dental fluorosis
Cross-sectional and longitudinal survey performed in 1985; follow-up to a cross-sectional
survey conducted in 1980 to test the effect of abrasion and remineralization on existing
fluorosis, and to monitor possible changes in fluorosis prevalence.
US/Illinois.; seven rural communities in Illinois; 432 children 8-10 yrs old and 193 children
13-16 yrs old, who had lived continuously in the same communities. Community water was
the primary source of drinking water for all children. Three cohorts: (1) 13-15 year olds in
1980; (2) 8-10 year olds in 1980 who are the 13-15 year olds in 1985; (3) 8-10 year olds in
1985.
None
8-10 yrs and 13-16 yrs. The 13-16 yr old children had been previously examined in a 1980
study (Driscoll et al., 1983). Developing teeth of Cohort 1 were at risk for dental fluorosis
from 1965-72; those of Cohort 2 were at risk from 1970-77; while those of Cohort 3 were at
risk from 1975-82.
The seven study communities were grouped into four categories based on the concentration
of fluoride in their drinking water; optimal, 2x optimal, 3x optimal and 4x optimal.
However, the F concentrations at each drinking water "optimal" category were not provided.
This information may be included in the 1980 reference survey performed in the same
Illinois communities (Driscoll et al 1983).
PROFILER'S NOTE: A later study (Selwitz et al 1995) of this series defines 1 ppm as the
optimal water fluoride concentration for the studied geographic area. Selwitz et al (1995)
also provides drinking water concentrations for the other areas, as shown below.
V.v,ii F.uor.il- CJIM uouvAi-ciPi-.
£*'ir~ t - 5V ~?i.!^t^lj *j CcCKa-itrul ^ Pi1"1"5 *^ ' 6T>' "L&V i"*1 t'«ri~!
OpLptiil i~ W Up *Vi * -ft. ^J ^ i ^i'* *X*^ C " »3rCT ^ i ,^™j^
F US SI "J
VJJITH, 1 IXvVU'n l 2HS w* M i*i 1.1 ">
r" 5\ 4^ It-
(Aon-:-*1 F W ?l
Source: Selwitz etal 1995.
Fluoride in drinking water was the only exposure route assessed quantitatively, although the
study authors noted that beginning in the early 1970's there were other possible sources of
exposure to fluoride, including commercial infant formula, other processed foods, fluoride
dentifrices, and fluoride supplements.
Methods for measuring the fluoride concentrations in water supplies of study communities
were not reported; other water quality parameters were not described.
Dental fluorosis and caries incidence in 432 children 8-10 yrs old and 193 children 13-16 yrs
117
January, 2008
-------
PARAMETERS
MONITORED:
STATISTICAL
METHODS:
RESULTS:
Dental fluorosis
old from seven communities in Illinois having different concentrations of fluoride in
drinking water (mean values ranging from 1.06-3.89 ppmin 1964-1980 and 1.01-3.59 ppm
in 1974-1990) were compared using acceptable methods of diagnosis (TSIF index for
fluorosis, and DMFS scores for caries). Methods of statistical analysis were not reported.
TSIF index (see Section 2) was used to evaluate fluorosis and DMFS scores (see List of
Acronyms) was used to evaluate caries. Only the TSIF index was used in the 1985 survey.
Not reported. Levels of agreement between examiners individual and combined scores for
all tooth surfaces were determined by Kappa values for 1980 (0.48 and 0.65 respectively)
and 1985 (0.51 and 0.64 respectively); these Kappa results indicate moderate agreement
between the 1980 and 1985 surveys.
IIII.IIIIIM ^
1 Table 2 • Pereentaf© distribution of TSIF scows for ill permanent tooth surfaces of 8
to 10 year olds by watef-flyeriete level In 1980 *nJ 1985, I
P^a mas^r' disirsliiiU^i; >X IS!* ^crm1?-
! ll-Sli
1
1 \\ ..!(-. [lil.iinK ,\',. ni
' (fiiiK j 6 > :': .1 i " -. 7
, *i|iu'ii.l i? dt; [•••> i:> if, 'iii ill iuB u ;>
>nj«i.Tut (•. 'ia it H,( o,!' ti.X !'- »'-' ''•!' iiO :
; '.",•(> mi;,l fc -8'' M.'li HI1- X.< l).'. M< i, ' iit '
1 4V "|iuai,.l 5)' TiT'.i S,si !?. ! it'.7 0! US '" i t ,
!!tS.r>
Ofuun sl i?ifi .-'J Jf».n S i- 18 l"i fi (11 C', (• "»•
j •_•• <>|yiuril Uly "-I! '.(H ill. H, ' Mil i. l.'J ;n>
'* i,|,ii",i, i !!'•' --sfi "if 1 ];;,> fl',; t\" ' *. ii. !i Mi
j-,.iV -M2 1:2 :^7 ;ft" Cirt ', i tt \ :i
TaMe 3 • Percentage distribution of TSIF scores for all permanent tooth surfaces
' for 1 3- to 1 5>year-oW children by water-fluoride lew»l in 1 98G and 1 9S5.
Pillt 111 JJ4C lllllilt.JUI.IBl. I rSlt'MUIli
ISSN
W. n !l',,ondc- So "jl
I nl .liiVtm Ii i » S 4 " i> ?
<>[ t .1* I.I w! h ill 1 ". s'.K Oil Ul. II'.) :i ',
;'• >un\il ;\" iil 7 i'ri ^ ', j< ,"..n iMi nJ iiw In
; '<"' \rtnrjjj M1 5 iil •!• ft '"" SlI '.'• j! 0.7 Hu ul
! <> ,,iiit,ji M .""..'t £''t". :t"7 ,i.f> r-'i \.;i n i -.i.;,
iw>
Opn'.,i; *! 7f 7 if <•• -I " ."H d 1 IM; fsu in
i L* e)jMiit',J! ^? >'-'.-"' 1- "> •*.'.' i*» l». ' l.S 00 >JU
1 s»uj,|iri.ii •!'< ''s'S q'i 9 '-x-J ^fi U "- 15 111 nfi
;• ij!
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE
NOT FOUND IN NRC
(2006)
PROFILER'S DMO
REMARKS 11/21/06
and
12/15/2006
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM.
LOEL/LOAEL
POTENTIAL
SUITABILITY FOR DOSE-
RESPONSE MODELING:
1 « M»n DMF turlnce $«*»* of children by age ea-iegc-ry and walW'ftuorld*
• ltv»f .n1SSO»nd 1985.
M,.!> «,.,.,.,, .!„
_•».. i.lf;
•,..-., -i '1 L ",i'!i N'" il Mt.ui IK !'>«.',> '-;">.> N.v. ..I M.sci;. '.Hl-mm i>, ->>",c, i ,
.,,,.' .l-.ii.hiii, j»\iPs I!.TTI >p-,,r,ai il..M|v, l>\ls-v ,:-.•«- .->ji;iT.;tl ir-Vii H* ''''' ' - -'-''•"
I ._. , _!,;,,.;,. ft! .:>: ,".',1 in-: i.-i: -.-' " '• *-*>
-. -.-.r.11,-1 x' "":" '">T "• i'i ^'*" "' " ' • '"
' ..v .,!,,-rji :<•. \-fi -i J ."", -i,Sr 's.'-' ; .
:. !•: :.= ••[,'!• ,.!.i
I'.,,: ,.,.,) !!l '• i •• 'M "' -Sl'ii
-.vjl.lhi "H! ' 'K >7 '» "" i "."• "fl.'l •'ill
; .»..., i-ns.1 >•! »' ."'•• -M "!•'! ;'! ' '!•'.- .
"
At 2X optimal fluoride concentration, additional intake from extraneous sources of fluoride
could be approaching a critical threshold for producing severe fluorosis. At 2X optimal,
7.6%of labial surfaces of maxillary anterior teeth of 13-15 years olds examined in 1985
exhibited severe fluorosis. The 13-15 yr ols showed little change between 1980 and 1985 in
the relative differences in the mean DMFS scores between the optimal and above optimal
fluoride areas. For to 8-10 yr olds, the mean DMFS scores in 1980 were substantially higher
at 4x optimal than at 3x optimal, but in 1985, there was very little difference between the
two groups.
Driscoll, W.S., et al. 1983. Prevalence of dental caries and dental fluorosis in areas with
optimal and above optimal water fluoride concentrations. J. Amer. Dent. Assoc. 107(1):42-
47.
Driscoll, W.S., et al. 1986. Prevalence of dental caries and dental fluorosis in areas with
negligible, optimal and above optimal water fluoride concentrations in drinking water. J.
Amer. Dent. Assoc. 113(l):29-33.
The recommended optimal fluoride level for the communities studied was reported in
Selwitz (1995) to be 1 ppm. The study authors note that, beginning in the early 1970's, there
were other possible sources of exposure to fluoride, including commercial infant formula,
and other processed foods, fluoride dentifrices, and fluoride supplements, but these were not
quantified. The use of bottled water by the study population was also not addressed. The
size of the study groups is sufficiently large for statistical analysis, and a standard method
was used by the study authors to evaluate dental fluorosis (TSIF index). The inter-examiner
reliability of the scoring system was also verified (Kappa values).
TBD
TBD
Not suitable (_)> P°or O> Medium (x), Strong (_)
Suitability medium to strong for the dental fluorosis endpoint. An adequate number of
119
January, 2008
-------
CRITICAL EFFECT(S):
exposure levels were considered and a sufficient number of children were examined.
Dental fluorosis
120
January, 2008
-------
Heller, K.E., Eklund, S.A., and Burt, B.A. 1997. Dental caries and dental fluorosis at varying water
fluoride concentrations. J. Public Health Dent. 57:136-143.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental caries
Dental caries and dental fluorosis in children aged 5-17 years
Data were obtained from the 1986-1987 National Survey of Oral Health of US
Schoolchildren conducted by the National Institute of Dental Research (MDR).
US: 40,693 children aged 4-22 years were included in the National Survey. The Survey
included at least 14 sampling strata, with two strata per geographic region. Only data for
children with a single continuous residence (N = 18,755 aged 5-7 for caries analysis; N =
15,532 aged 7-17 for fluorosis analysis) were considered.
No control population was included.
Only children with a single continuous residence were included in the current analysis;
children who had resided at more than one address at some point in their life were
excluded.
The fluoride level in the school drinking water was used as a measure of exposure. For the
analysis, fluoride levels were presented as <0.3 ppmF, 0.3 to <0.7 ppmF, 0.7-1.2 ppmF,
and>1.2ppmF.
The exposure assessment consisted solely of measured fluoride concentrations in the
child's school drinking water. A written questionnaire, administered to parent or guardian,
included questions regarding the use of fluoride drops, fluoride tablets, professional topical
fluoride treatments, and school fluoride rinses; these findings were used for statistical
analysis only, not exposure assessment.
The method for analyzing fluoride in the drinking water was not described.
Little information was given in the current publication. Design and conduct of the original
survey have been described by US Public Health Service (1989, 1992) and Brunelle and
Carlos (1990). Data for the current analysis were obtained from a public use data tape
provided by the NIDR. Comparisons between at least 5 field examiners found good
agreement on paired t-tests.
Oral examinations of children included visual and tactile assessment of dental caries and
restorations using the diagnostic criteria of Radike (1972); no radiographs were taken.
Children in grade 2 and higher were examined for dental fluorosis. A classification system
based on Dean's Fluorosis Index (Dean 1942) was used to evaluate all erupted permanent
teeth. Fluorosis prevalence was determined by whether or not the child had at least two
teeth with a score of 1 (very mild) or greater. A written questionnaire, administered to
parent or guardian, included questions regarding the use of fluoride drops, fluoride tablets,
professional topical fluoride treatments, and school fluoride rinses.
The Statistical Analysis System (SAS) Version 6. 10 (16) was used for data management
and for descriptive statistical procedures. The SUDAAN (Survey DAta ANalysis)
Release 6.40 statistical program (17) was used for statistical tests because of the need to
adjust variances for the complex sample design of the NIDR survey. Sample weighting
to represent the population of US schoolchildren was used for all analyses.
Mean scores for both decayed or filled surfaces (dfs) in children aged 5-10 years and
decayed, missing, or filled surfaces (DMFS) of permanent teeth in children aged 5-17 years
decreased with increasing fluoride levels (Tables 1 and 2). The mean dfs score for the 0.7-
121
January, 2008
-------
1.2 ppm F group was statistically different from the <0.3 ppm (p=0.004) and 0.3 to <0.7
ppm (p=0.045) groups. Statistical significance was also found between the >1.2 ppm
groups and the <0.3 ppm groups (p=0.03 1). For the permanent teeth, the mean DMFS
score for the 0.7-1.2 ppm group was significantly different from the <0.3 ppm group
(p=0.003). No other statistically significant differences were found between DMFS scores.
TABLE 1
PistribtstKm and Meais of dfe Scares fcy WaitrFlwdde Slate*
d(s(%J
nt N%| 0 1-5 6-10 11-20 i20 Mean (SO
«03ppmF i«2 38-9 *>-S 25.7 12.6 11.1 3.9 4.49(0.28)
03-eO.?ppmF 1,035 8,3 45.4 293 12.1 10,1 3.1 4.18 {0.27}
0.7-1 .2 ppm F 420S 49,8 51,1 27.9 11.1 7.6 2.3 3.33(0.23)
>1.2ppmF 415 3,0 SOA 28& 11.5 7.9 1.7 3.42(039)
Ail 9,77? 100 48.9 27.1 11.8 9.1 3,0 3.S?{0,17>'
*Sc«ts are sr^fidMdixed to the age and $« distributor OS" US school Aildr«m aged 5-10 years who hsd a history of a sii^te 'resi,4enM,
iSampk §1?^,
JWaghled potwl»tion percentage,
TABLE 2
Distribution and Mean of DMFS Scores by Wal«t Fh»ndc Status*
DMFSCSO
nt K%J 0 1-- fr-10 n-:0 >20 MeanlSS)
<0,3ppmF 7,584 3O S3.2 258 12.5 5.6 1.9 3JB(0-!5)
0.3- <0,7 ppm P 2,113 10.1 S7.I 2~>-9 12.2 S.4 1,3 2.71(0.12)
0.7-l-2ppmF 8,097 50.4 557 271 11.8 5,0 ' 0.8 253J0.1H
>1.2ppmF »91 3,2 525 23 C 9.8 8.1 ' 0.6 2.80 (039)
Ail 18,755 100 54,6 26,3 12.1 5.7 1.3 2.7S 10,09)
"Scores art sIMdwetesw! to the ag<» and s@x disftifoufiesn of IIS schcwkhi^drra ^gcd ^.17 y*as who had a history of a Mpgl^reidenee
tSample «e
fW«gbt«l papulasion |>erc«t*fage,
Additional graphical comparisons and statistical analyses were made between caries levels
(dfs or DMFS) and fluoride water levels. As depicted graphically (Fig 1), both dfs in
primary dentition and DMFS in permanent dentition decreased between 0 ppm F and 0.6-
0.7 ppm F, then plateaued up to 1.2 ppm F; dfs continued to decrease at higher F
concentrations.
122
January, 2008
-------
Dental fluorosis
FIGURE 1
DMFS for Children Aged 5-17 Years, and dfs Scores for Children Aged 5-10
Years, by Water Fluoridation Level for US Schoolchildren with a History of a
Single Residence? Scores are age- and sex-standardized to children with one
residence aged 5-10 years for dfs and aged 5-17 years for DMFS.)
dfs or
6
4.
2.
•1 .
DMFS
/\
/ V"-\
L N^^— *-— ^/\ ^ i
^o^- -• ** *
^^ Vo^ 7
^ V"
0 0,2 0.4 OJ OJ 1
ppm F
f —-dfe — -0MFS
1,2 1.4 IS*
Multiple regression analyses were made between dental caries levels with demographic and
fluoride exposure variables. In primary dentition, a lower dfs score was significantly
associated with being female, higher water fluoride levels, and for having ever used
fluoride drops. In the permanent dentition, lower DMFS scores were significantly
associated with increasing fluoride water levels and the reported use of fluoride tablets. In
contrast, increasing age and being female were associated with increasing DMFS scores.
Mean fluorosis severity increased with increasing water fluoride level (Table 5). The mean
severity score for the >1 .2 ppm F group was significantly greater than all other groups
(p=0.045-<0.001) and the mean score for the 0.7-1.2 ppm F group was significantly greater
than that of the <0.3 ppmF group (p=<0.001).
TABLE S
Dhtriljulion and M*an of Fluorosis Severity Scows, and Huoiosw Prevalence, by Witei Hueridatian Status*
FltsoTC&is Seventy ^%\
nt IV 74 .1 05 1 2 3 4 (Si) (SF1
•dOppir, F 6,!3<) J5.2 5«J 26 to 10,7 24 04 0.1 0.30(0,03) 11 Ml")
03-<.«)7ppmF 1,?«1 104 474 310 173 31 12 ft.C 0.43{O.B) 21 "'(60)
Q,7-l,2ppmF b?28 51.! 336 365 22.5 5.8 1.3 00 OJ8(O.CB) 291(34)
M2ppmF 772 33 28 1 305 27.2 70 53 2.0 0.8010.10) 414(44)
Att '5,532 100 441 323 !7,9 43 1: 0,3 0.47(0.04) 23S(2tt
•Sans iai standardised to Ibe age and sex cfatababon of US schoofchildj-pn aged 7-17 ytara who had a Maory of .; single reidcnu
{Weighted popubbon perantige
Sffavmg it least swo («h with Dean 5 fluoroas icons 1 (very mild) or greater
Graphical representation showed both fluorosis prevalence (%) and severity increased with
increasing water fluoride levels but decreased with increasing age (see study Figs 1 and 3).
123
January, 2008
-------
Dose response
FIGURE 2
Fluorosis Prevalence and Mean Severity Scores by Water Flaoridation Level for
US Schoolchildren Aged 7-1.7 Years with a History of a Single Residence
(Scores are age- and sex-standardized lo children aged 5-17 years with a history
of a single residence, Flaorosis prevalence was defined as having two or more
teeth with veiy mild fluorosis or greater; mean severity scores were derived in a
manner stoUmr to Dean's CH scores.)
Fiyorosls Ftaorosis
prevalence (%} severity
1
f?\
?
/\ jfc
r^v \^
•=-^Tyv-/ "
- *t
- n n
• n K
0™
0 0.2 0,4 0.6 Q.S 1 1.2 1.4 1.6+
ppm F
\ —Fluorosis prevalence (%} — *-Flyorosis severity ]
FIGURE 3
Huorosis Prevalence and Mean Severity Scores for US Schoolchildren Aged 7-17
Years by Age with a History of a Single Residence
(Ftaorosis prevalence was defined as having two or more teeth with veiy mild
fluorosfc or greater; mean severity scores were derived in a manner stailw to
Dean's CH score.)
Fluo
p re vale
30 •
25
20 -
11 -
5 -
0_
rosis F
nee (%)
^^^_ __ ^
^ " X^~^*vs^
^V*—
luo rosis
severity
0.6
o.s
0,4
0,3
0,2
0,1
7 S S 10 11 12 13 14 15 16 1?
Age
| — * — Fluoresis prevalence {%) ••* -Flyerosls severity |
A logistic regression model, controlling for age and fluoride product use, showed children
who consumed water at 0.3-O.7 ppm F, 0.7-1.2 ppm F, or >1 .2 ppm F had an odds ratio of
2.07 (95% CI = 0.92, 4.67), 3.32 (2.25, 4.91), or 4.96 (2.87, 8.58), respectively, of
developing fluorosis compared to children who consumed water at <0.3 ppm F.
Increasing fluoride water concentrations were associated with lower dfs and DMFS
scores up to a level of approximately 0.6-0.7 ppm F. Higher fluoride water
124
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date:
REMARKS CSW/1/3/07
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
concentrations did not affect DMFS scores while dfs scores showed further decline at
>1.2 ppmF.
In contrast, increasing fluoride water concentrations were associated with higher
prevalence and severity of fluorosis with no apparent plateau of effects.
PROFILER'S NOTE: No attempt was made by the authors to estimate intake of fluoride
from drinking water consumption; thus, a dose could not be determined.
A suitable trade-off between caries and fluorosis appears to occur around 0.7 ppm F.
Little decline in caries levels were observed between 0.7 and 1.2 ppm F, while
considerable dental fluorosis was seen at this water fluoride level.
US Public Health Service, National Institute of Dental Research. Oral health of United
States children. The National Survey of Dental Caries in US Schoolchildren: 1986.1987.
National and regional findings. NIH pub no 89-2247. Washington, DC: Government
Printing Office, 1989.
Brunelle JA, Carlos JP. Recent trends in dental caries in US children and the effects of
water fluoridation. J Dent Res 1990; 69 (Spec Iss):723-7.
US Public Health Service, National Institute of Dental Research. Oral health of United
States children. The National Survey of Dental Caries in US Schoolchildren: 1986-1987.
Public use data file documentation and survey methodology. Washington, DC:
Government Printing Office, 1992.
Radike AW. Criteria for diagnosis of dental caries. In: Proceedings of the conference on
the clinical testing of cario static agents. Chicago: American Dental Association,
1972:87-8.
Dean HT. The investigation of physiological effects by the epidemiological method. In:
Moulton FR, ed. Fluorine and dental health. Washington, DC: American Association for
the Advancement of Science, 1942, pub no 19:233 1.
This study was an analysis of data collected as part of a larger national survey. The authors
assured consistent water fluoride exposure by excluding those children who had resided at
more than one address during their lifetime.
Only school water fluoride levels were measured; no attempt was made to correlate those
levels with home water supplies. It might be assumed that municipal water supplies would
be similar between home and school, but this is not true for rural homes which might use
well water or cisterns.
Despite lack of data on other potential fluoride sources, the study showed clear correlation
between increasing school water fluoride levels and decreasing caries with concurrent
increasing fluorosis. Severe fluorosis (Dean score of 4) was observed in 2% of the
individuals in the >1.2 ppm F group; however, a correlation of fluorosis and a significant
increase in caries was not evaluated for individuals in this group.
Doses could not be reconstructed because these data were not collected.
Could not be determined.
Could not be determined.
Not suitable ( ), Poor (X), Medium ( ), Strong ( )
125
January, 2008
-------
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
A positive correlation was found between increasing water fluoride levels and decreasing
caries and increasing fluorosis; however, the data were combined into concentration ranges
and therefore may not be suitable for statistical analysis.
Dental caries, fluorosis
126
January, 2008
-------
Hong, L., S.M. Levy, J.J. Warren, B. Broffitt, and J. Cavanaugh. 2006a. Fluoride intake levels in
relation to fluorosis development in permanent maxillary central incisors and first
molars. Caries Research. 40:494-500.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
Dental fluorosis (central incisors and molars)
US/Iowa: Prevalence study of dental fluorosis as part of longitudinal study of daily fluoride
intake in male and female children recruited at birth (Iowa Fluoride Study; see also
companion report Hong et al 2006b). Study conducted March 1992 to February 1995
included exam of early erupting permanent teeth at mean age 9.3 years.
Children (319 males and 309 females) aged birth to 36 months included in the Iowa
Fluoride Study; total fluoride intake estimated from parental questionnaire administered
every 3-4 months, and dental fluorosis measured in early erupting permanent teeth at subject
age 8-10 yrs. The cohort was 98% Caucasian, from families with a relatively high
socioeconomic status (71% having a family income of $30,000 or more and 46% of mothers
having completed 4 years of college), 44% were first children, 32% were breast-fed for at
least 6 months, 4% had low birth weight (<2,500 g) and 3% had developmental disorders.
Children were entered into the study using Institutional Review Board-approved consent
procedures.
None
Birth to 36 months (and beyond)
Estimated daily intake was categorized into three categories: O.04 mg F/kg BW, 0.04-0.06
mg F/kg B W and >0.06 mg F/kg B W based on questionnaire analysis.
Fluoride intake from water, beverages, selected foods, dietary supplements and dentrifice
was estimated from questionnaires (unvalidated) completed by parents every 3-4 months
from subject birth to age three years. Fluoride intake (mg/kg BW per day) for cumulative
time periods (AUC) was estimated for 1st year (0-12 months), 2nd year (12-24 months), 3rd
year (24-36 months) and composite all 3 years (0-36 months).
Study methodologies described previously in Levy et al (2001, 2003)
The purpose of the study was to report the fluorosis prevalence by levels of estimated daily
fluoride intake. As part of the Iowa Fluoride Study, children were followed from birth to 36
months with questionnaires every 3-4 months to estimate daily fluoride intake (mg/kg BW)
from water, beverages, selected foods, fluoride supplements and dentrifice. At 8-10 years
(mean 9.3 yrs), 628 children (3 19 males and 309 females) were examined for fluorosis on
permanent incisors and first molars using the Fluorosis Risk Index (FRI; see NRC, 2006,
page 90). Fluorosis was differentiated from enamel demineralization ("white spot") based
on color, texture, demarcation and relationship to gingival margin. The FRI was adapted to
include assessment of all visible enamel surfaces, with four zones scored separately on each
buccal surface (the incisal edge/occlusal table, the incisal/occlusal third, the middle third
and the cervical third).
Dental fluorosis on early-erupting permanent teeth (8 incisors and 4 first molars) was
determined at 8-10 years of age (mean 9.3 years, range 7.7-12.0) using the FRI. Fluorosis
was differentiated from non-fluorosis opacities by the criteria of Russell (1961). Fluorosis
was further differentiated from enamel demineralization ("white spot") based on color,
texture, demarcation and relationship to gingival margin. The FRI was adapted to include
assessment of all visible enamel surfaces, with four zones scored separately on each buccal
surface (the incisal edge/occlusal table, the incisal/occlusal third, the middle third and the
cervical third). Scoring criteria differentiated no fluorosis, questionable fluorosis (less than
127
January, 2008
-------
50% of zone with white striations), definitive fluorosis (greater than 50% of zone with white
striations) and severe fluorosis (zone displays pitting/staining/deformity). Cervical zones
were excluded from the analyses due to lack of consistent full eruptions. Incisor fluorosis
was defined as having FRI definitive/severe fluorosis (FRI score 2 or 3) on both maxillary
central incisors. First molar fluorosis was defined as having definitive/severe fluorosis on at
least two first molars.
STATISTICAL METHODS:
Correlations among fluoride intakes for the first three years were assessed using Spearman
rank correlation analyses. Logistic regression was used to assess the relationships between
estimated fluoride intakes and fluorosis. Fluorosis prevalence rates were calculated by
fluoride intake category and relationships were assessed using Cochran-Armitage tests for
linear trends using scores equal to the median fluoride intake for each group. Using fluoride
intake of O.04 mg F/kg BW as a reference group, the relative risks for fluorosis were
calculated for 0.04-0.06 mg F/kg B W intake and >0.06 mg F/kg B W intake. Similar
analyses were used to compare subjects who consistently stayed in the same fluoride intake
category for the first two years and those who were consistent for all three years of intake
monitoring.
RESULTS:
Most dental fluorosis observed attained a FRI score of 2 (mild to moderate); <1.5%
exhibited a FRI of 3 (severe fluorosis).
Results of the study are shown in Figures 1 and 2 and Tables 1, 2 and 3 directly from Hong
(2006) (OR = odds ratio):
•s
e
i
•••••• •"" Age 0-1 Cm =1.103 — Age 2-3 COR = 1 .1 OH
/^
^ '
sf
.,-•
002 O.04 0.06 0.08 O.1O 0.12 0.
Fluoiide intaK* t,mg R &g BW»
4
Fk| 1. rri"bitbilityfifj,v_Tjn;i nfiit maxillary >~taiti'jl inciv,tt tlu«Tn-
r.is cast b)' level i. it'flu- "fitte inlu.Ue iff nt Jughlic i>jgre:.Mi'n) ' »R«.
t'.u' incr«/mtnts t«t'f ni my P-'kg BW. p valut1. fur t'ltu -nek- uiUiJ-"t>
'
128
January, 2008
-------
Table 1. Prevalence
iS^t^.^V:-,,
ISpi-;
0.8-
g- Of,-
.0
a- 0.4-
02-
i
Age 0-3 OR = I 48j - *g« 1-2 i OR = 1 2Sj
- Age 0-1 (OR = 1 U7> Age 2-3 iC'R = 1 3-»,i
-
. * " ""
"i 002 004 00t> 008 niO 012 OH
Flu' snde mtatte- tnig F Ig B\^*
Fli/, 2. rrt'txituhtyi if permanent first ni- •larfluiipv.K-cjvL hylev-
fl • '{ flui •ridtintake.1 tin mi li >(;txtic regie:.*-li -o " OR*, fi jr incremt.'nts.
_«f ' M '1 mg F.'lru, E\\r p volu.es (• U" fluoride int«i b."» were ,ili itatisli-
rallv sigrsit" cant ip< O.U5I
Df flmstosis ^m bofife pensianent fflastllary cgiatral incisors by esttiuafed tola! lluorMe intake
U^i'^Vf
'. v:;:,;,,.f;t.l ;
0-12«montt« 405 185
12-24 months 405 in
24- 36 months 405 136
O-Kmondui 485 132
2 years steady3 282 121
JyesrssteMly4 113 67
^'iiS(iji^i(^^i4^^
/^A i ,'?O»-'>'v,ijkvA:-;fc>, 'v ;tt',i|; <•;,;' Jifv-'' ^W-^^'M:,^
-^ttM<«; .•'.>"'i,; ,:;;;.,jh«**^«tji,;*5';Vyv>:: o ;,;';•';*?%!'. 'ffftj
•;!';=::Xv"?*^
IBs-SS
1S.7 67 2S.4 l,S2(0.95,i75) 153 32,7 2,08 (1J9, 3,12)* QMl
16J 144 27.8 1,70 (1.12, 2Mf 83 32,5 2.00 (1.27, 3.1S)1 0,002
ISA 148 19.6 1.87(0.66,1,73) 121 347 1.19 (1.23, Z»)J 0,002
12.9 165 23.0 1.79 (1.0«, 3.02)2 108 38,0 Z95 (1.71, CSg)1 O001
12.4 32 2J.1 2,27(1, 09, 4.70)2 49 «&9 3,79 (116, 6.S3)5 OJ001
13.4 16 25.0 1.86 (0.68, 5.29) 38 50,0 3,72 (1.84, 7.S4)2 Q.001
3 OscferaM- Armitage fest for linear trend.
1 RH sigHitk»n% greater thin 1.0 (p < 0.05) when compared to group wtth «Q.Q4 mg BkgMaf.
3This |K»p hid the same Intake category (<0.04, 0 J4-0.06 or >i.06 mg WkgMay) far fears 1 and 2.
4 This group had the same Intake category (i.06 mg F/kgMay) for years 1, 2, and 3.
Table 2. Prevaknce of f hjorosis on two or more permanent first mokrs by estimated total fluoride intake
Fluoridf tetib
period
si*[«ti 0.06 mg Rkg/day! far years 1 and 2.
*This group had the same intake category (<0.04. 0.04-0.06 or >0.06 mg F/kg/day) for years 1, 2, and 3,
Table }, FhiC'iTrsb CJMJ prevalence h
0- to 12-month
daily F intake
12- to .Vi-mont;
.0.l36 ". . 1 2 r.;5 .? j 7 1 % 1 3 >'4 2'i 1 9 '">"<.• ! 1 1
7 >0.1'0
5 20.5°oiS'39)
J) 4i.2'V., (7*'I71
•5«i J«frtu (2 1/53)
F.«ch '.*(' da\otcr. ,s sithjaxt with were iil«I= y< llti-jri-Us, AH -verier Hibk'Cts with (luon^:; wrE~eratodnHldto o^^deratf f!-B! = 2K
129
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'
S
REMARKS
Initials/Date
VAD/12-26-
06
PROFILER'S ESTEM.
NOAEL
PROFILER'S ESTEM.
LOAEL
SUITABILITY FOR DOSE
RESPONSE MODELING
CRITICAL EFFECTS:
The fluorosis prevalence rates increased with increasing fluoride intake levels, although the
rates varied substantially across different time periods. Subjects with daily fluoride intake of
<0.04 mg F/kg BW had less than 20% probability of developing fluorosis and almost all
was of mild severity (FRI score 2). Daily intake of 0.04-0.06 mg F/kg B W had a significant
elevated risk for fluorosis while daily intake of >0.06 mg F/kg BW was associated with high
risk of fluorosis on early-erupting teeth. However, the data showed different susceptibility
for fluorosis by tooth type with maxillary central incisors having greater fluorosis
prevalence than first molars. Hong et al (2006) consider that fluorosis development relates
to not only stages of enamel formation, but also to duration of fluoride intake level.
Hong, L., S.M. Levy, B. Broffitt, JJ. Warren, M. J. Kanellis, IS. Wefel, and D.V. Dawson.
2006b. Timing of fluoride intake in relation to development of fluorosis on
maxillary central incisors. Community Dent Oral Epidemiol 34: 299-309.
Russell, A.I. 1961. The differential diagnosis of fluoride and non-fluoride enamel opacities.
J Public Health Dent 21:143-146.
Since the cohort was 98% Caucasian and 7 1% were from families of high socioeconomic
status, results are not representative of the general US population. Individual fluoride intake
estimates varied with approximately 50% of subjects staying within the same category
(O.04, 0.04-0.06 or >0.06 mg F/kg/day) for two years and 28% in the same category for
three years. No data on the fluoride intake levels were supplied for those children with
severe fluorosis (FRI score of 3). In addition, there was a discrepancy in the study report as
to how many children had severe fluorosis. The first sentence of the Results section states
that 8 individuals (1.3%) had severe fluorosis (FRI score of 3), whereas the Discussion
section refers to 6 subjects (approximately 1.5%) with severe fluorosis (FRI score of 3).
Fluoride intake was estimated for the first three years of life, whereas fluorosis evaluations
were done at 8-10 years. The contribution of fluoride intake for the five years that intake
estimates were not performed (between 3 years of age and date of dental fluorosis exam)
was not addressed.
This raises issues related to the window of vulnerability.
The study design did not identify a no-fluorosis intake dose. The probability of fluorosis on
either the maxillary central incisors or permanent first molars at each dose category was
calculated. All but 6 or 8 of the 628 children (see Profiler's Remarks) had mild to moderate
fluorosis (FRI score of 2). No data on the fluoride intake dose categories for the children
with severe fluorosis (FRI score of 3) were provided.
The variations in fluoride intake over the three years of study estimates complicated the
establishment of a LOAEL. Only 1 13 of the 628 children in the study were in the same dose
category for three years. The influence of this fluctuation, especially increases in fluoride
intake, was not assessed.
Not suitable,Q; Poor (_); Medium (X); Strong (J
Dental fluorosis (central incisors and molars)
130
January, 2008
-------
Hong, L., S.M. Levy, B. Broffitt, J.J. Warren, M.J. Kanellis, J.S. Wefel and D.V. Dawson. 2006b.
Timing of fluoride intake in relation to development of fluorosis on maxillary central
incisors. Community Dent Oral Epidemiol 34:299-309.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION
STUDIED:
CONTROL
POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT
ANALYTICAL
METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
Dental fluorosis (maxillary central incisors)
US/Iowa: Prevalence study of dental fluorosis as part of longitudinal study of daily fluoride
intake in male and female children recruited at birth (Iowa Fluoride Study; see also companion
report Hong et al 2006a).
Children (297 males and 282 females), age 8-10 yrs (mean age 9.2 years), included in the Iowa
Fluoride Study. The cohort was predominately Caucasian, generally healthy, and from families
of relatively high socioeconomic status (see Hong, 2006a, for specifics). Children were entered
into the study using Institutional Review Board-approved consent procedures.
Control group (n= 18 1) included children from the cohort without fluorosis on either of the
maxillary central incisors.
Birth to 48 months (and beyond). Study conducted March 1992 to February 1995, included
examination of early erupting permanent teeth.
Mean daily fluoride intake over 48 months was 0.053 mg/kg/day (range 0.045-0.062 mg/kg/day)
in children exhibiting fluorosis and 0.043 mg/kg/day (range 0.038-0.049 mg/kg/day) in children
without fluorosis.
Fluoride intake from water, beverages, selected foods, dietary supplements and dentifrices was
estimated from questionnaires (unvalidated) completed by parents every 3-4 months from subject
birth to age four years.
Study methodologies described previously in Levy et al (2001, 2003).
The purpose of the study was to establish the relationship of fluoride intake during the first 48
months of life with fluorosis on early-erupting permanent teeth. As part of the Iowa Fluoride
Study, children were followed from birth to 48 months with questionnaires filled out by parents
every 3-4 months to estimate daily fluoride intake (mg/kg BW) from water, beverages, selected
foods, fluoride supplements and dentifrices. Fluoride intake (mg/kg BW) was estimated using
means (standard deviation), range and percentiles for individual time periods (four months) and
for cumulative time periods (0-12, 12-24, 24-36, 36-48, 0-20, 0-36 and 0-48 months) by the area
under the curve (AUC) trapezoidal method. The estimated daily average fluoride intake was
categorized into tertiles (low, middle and high fluoride intakes) based on the frequency
distribution of average fluoride intake for each of the first 4 years separately.
At 8-10 years (mean 9.2 yrs), 579 children (297 males and 282 females) were examined for
fluorosis on both maxillary central incisors using the Fluorosis Risk Index (FRI; see NRC, 2006,
page 90). Fluorosis was differentiated from enamel demineralization ("white spot") based on
color, texture, demarcation and relationship to gingival margin. The FRI was adapted to include
assessment of all visible enamel surfaces, with four zones scored separately on eachbuccal
surface (the incisal edge/occlusal table, the incisal/occlusal third, the middle third and the
cervical third).
Fluorosis was determined using the Fluorosis Risk Index (FRI; see NRC, 2006, page 90).
Fluorosis was differentiated from non-fluorosis opacities by the criteria of Russell (1961).
Fluorosis was further differentiated from enamel demineralization ("white spot") based on color,
texture, demarcation and relationship to gingival margin. The FRI was adapted to include
assessment of all visible enamel surfaces, with four zones scored separately on eachbuccal
131
January, 2008
-------
STATISTICAL
METHODS:
RESULTS:
surface (the incisal edge/occlusal table, the incisal/occlusal third, the middle third and the
cervical third). Scoring criteria differentiated no fluorosis, questionable fluorosis (less than 50%
of zone with white striations), definitive fluorosis (greater than 50% of zone with white
striations) and severe fluorosis (zone displays pitting/staining/deformity). As many cervical
zones were incompletely erupted and not able to be scored, three zones (incisal/occlusal edge,
incisal/occlusal third and middle third) were used in the main analyses. A fluorosis case for
regression analyses was defined as having FRI definitive or severe fluorosis on at least one zone
of both maxillary central incisors; controls had fluorosis on neither of these incisors. Subjects
with only one maxillary central incisor with fluorosis were excluded. Subjects with only FRI
questionable fluorosis were grouped as questionable fluorosis, unless they were excluded
because the required three zones could not be scored due to reasons such as incomplete eruption.
Subjects with questionable fluorosis were not included in the analyses.
Fluoride intake (mg/kg B W/day) was estimated using means (standard deviation), range and
percentiles for individual time periods and for cumulative time periods by the area under the
curve (AUC) trapezoidal method. The differences in fluoride intake in mg/kg B W between cases
and control were assessed using two-sample t-tests first at the surface zone level and then for
combined zones (incisal/occlusal edge, incisal/occlusal third and middle third). The correlations
among fluoride intakes for the first 4 years were assessed using Spearman rank correlation
analyses. The estimated daily average fluoride intake was categorized into tertiles (low, middle
and high fluoride intakes) based on the frequency distribution of average fluoride intake for each
of the first 4 years separately. With fluorosis defined as having FRI definitive or severe fluorosis
on at least one zone (incisal/occlusal edge, incisal/occlusal third and middle third) of both
maxillary central incisors, the relationships between fluoride intake of individual years and
fluorosis were assessed using logistic regression analyses. The odds ratio (OR) and
corresponding P-values were calculated. Akaike Information Criteria (AIC), a measure of lack-
of-fit, was used to assess the fit of the model. Based on the -2 log likelihood estimate, AIC adds a
"penalty" for each parameter in the model which offsets the decreased lack-of-fit associated with
models using more parameters. Thus, the AIC can be used to compare singe-parameter models
(lower AIC is preferable) as well as to compare models with differing numbers of parameters.
Generalized R2 values were used to examine the predictive power of logistic regression models.
Fluoride intakes, based on the tertiles during each of the first 4 years of life, as individual
predictors of fluorosis on maxillary central incisors and whether these variables remained
significant after controlling for other years was tested. The two-way interactions among
individual years were assessed; the significance was set at a = 0.05. Receiver operating
characteristic (ROC) curves were used to assess the relationships between fluoride intake (mg/kg
BW) and fluorosis during the different years. A ROC curve is a plot of sensitivity versus (1-
specificity) for each possible threshold for the predictor variable. The sensitivity and specificity
values were computed for each subject's yearly fluoride intake and saved as output to construct
the ROC curves.
Results of the study are shown in Tables 1 through 5 and Figure 1 directly from Hong (2006b):
132
January, 2008
-------
table 1. Fluoride intake distribution (mg;
\$c iV Mean (SU)
Individual periods (months)
Birth to 3 559 0.055 (0.056)
>3 lo 6 545 0.057 (0.0471
>6lo9 564 0.054 (O.(Hl)
>9 lo 12 55s* 0.040 (0.030)
>12 lo 16 533 0.041 (O.IE7)
>16 to 251 52H 0.051 (O.IE9)
>20lo24 551 0.052(0.031)
>24 lo 2« 542 0.050 (O.IE9)
>2H lo 32 541 0.052 (O.(EH)
>32lo36 421 0.052 (O.IE7)
>36lo40 336 0.052(0.032)
>40lo44 313 0.047 (O.(ET)
>44to4H 396 O.(t44 (0.029)
Cumulative periods (tnonth.sj
0 12 514 0.052 (0.036)
12 24 44CI 0.046 (O.IE3)
24 36 444 !S.!!52 (O.(ES)
36 4H 430 0.049 (0. ','li
• I*iti4'> !'2r
•4, l,i 44 ' -4
•44 tn ^ ;> '4
Oitnulilivf fWiiid- iti'iiiith^i
I In ^ * !(
1 :2 • li'~
•'2 In 24 >\M
•24l,i >i • ' '"
•Vtn 41* it ^
"l,i 2' 't'""
! Ui ¥ • " t,
,11,, 4-* '!4,"
"lln1 mitrli't- >)l Hibjivl- wli,i r 'inrtitsJ 1)1
"1 h ' mill In is Hi i'.M1- itid iit'iK 3-s'- \ sn
,\iutrti
I'iuut.hi^iiii Uitli ti'Hilhn , I'nlril tni.t-,1
jtl bid
Ihitd-
2H
Y2
- ' 'ii
it '"
J* '
,U']
\r
<\2'
ii2«
Kl;
il-*
17,
; i*
•i i;
>'r ' ^
',' •;
" '2
1 1,
' i'
{} r
i,! 'N
Kiltiitiii on"-
'j Uu f.ivh
-, t.ltl-ldt'tlt
MidJk-
Ulttds
* it
n™
'i4'
!1",
' '2
[i '
i| W)
') >i
i |4
Kd
ii 1
I«
•1 "i
U'2
'i-,'1
i,fl
i n
1 IH
':4
li
12'
ir
virwdl.itd
rmi
l-Kl
4>J
4*
\' 2~
T2"
i' '•,
("«'
j.1 Tl
!' !^
i1"^
r*
c:4
1 n
i'4*
'' >
>' 12
('27
l' |*i
1'4''
l' "
r' '1
i' V
I'.M
ttfi-j«.l n-w»it
ittlun'ii, Jt'ft'tidtdi,; ifi tin
,; Ihnv MIH
• itid t'uiudit
il'tl Sin.lFO-4- , l-V*
Itid tl D) ^-l>
III
tllltuLlliU"1 till V fVIL lj>-
'i! lhirdv
mt v:i
,
,
.
V
.'
,
H; ttirt1 fVti.id' in
, w ind ..ifilt.tf
i; .vi \itjl , .Air*
\m fMi< ,ith
''tilfjt iaii-.it
^
*
! ^
< ,^
! 'r
: 3
!',':
; 'i
» * j
ir
!•':
'»?
'^;
* ;,
V
1 ';
! .n
• '
? i
S 1 '
'J
1 2
I'ktl U llUIfll
li'titnli,in Kir
ilii
til,'
133
January, 2008
-------
lilWi'"!. Mo/in .IsiK thmriJc inliiki ahJ
t\W \
'fiJtuJUil } wiiitK liTunlli i
Bittli In J ^
•It.ift >'«
-f la •< "i W
••:2L.i- il
•lMn 2,; 2^
•2'} Ui 24 W
-• 24 1,12* It,'
>2-li, <2 2«
SiT: is"
•4:i»i44 "^
••4*l,i 4^ 224
I'lJtl ulltlVC f-H'llilil- ItlUitltll^l
,' ",2 2~*
•:2l,i 24 2^
•24 1<- *i 2>
i,. i S ji 1*11!
„* -^ |- 1 ^,*j ^ i^
' ki 2.' 21*-
' In > >4
t,, 4-* ^
Huiircj-i* ,in b.ith MM MlLit t i uitrai iivt;
l'.,,i -M pic 1 Ir4 ti'Hiit;
K,i tlu ns^t"1 With llu.'M-i- )' Vakii1
i-i"ir til. it n tr, l» i4*"1 CIi, tri; ki; tn,' if t.M J
,' 4~|"(1^4 'I '"TI i\tfr-ri |i" '"51 •' '~4l '* ".
'ii4ri"'t4"' 'ir^ii i*'r; iiii'ft, ,% 7-41 - • "
' P41"!,' Hi '.!. 'Wi i'.i>-l K '-"ihi1"",! ..',.'
' T-ri.'iT1-! I '4 i '1 "'44 nl ''4^ " '4ii ' '"
1 , '4" i,! ('42 '1 VD I'iFi" iij 'i2 * '"11 * ' '1
,V4" i ' 114; ,v ",'"i » ifc'a |iVi-,i .I'H'I •• •,
\'4"i • 141 ,' u^i i|ipa.(ih14') t'.^i ',"
1 J14~(,',iI4) ,' ,'n i (UV'tit'i^S '1.':;iJt ','""
'i'4>.|('v4 l\14-ii iii> ' n'i'.4 ,'ptN) .^^M
: !4Ti!42 >u» ',',-1) >'i>-i I'K.VM' >*i • ,' l
','4; i, j'17 i;.,'44i il-Fl HI "47 ,' '^i • ' '"
,' I4H i '44 ,'"^1 d>B7 i'i'x!?2 '''h2i
' '441, '-14' i-V^i il-F) n,i',;4« V~St -''
,' '-t1!, 'H-i 't I47i !liIl;H i'1'Til >\'*-4i - ' '1
;i'4t;i;r4' .; ^ ihp-i n!i'"4 '.JVi
'•t!4iri|N -J.,'4^1 | >R1h'''4" ,1 "-2I ,* "2
ft - .'niistJi'ntm thd'i1 .men, excluding cervical zones.
134 January, 2008
-------
I^b!«l4 I.ii;i.lk nudi'l
fliwnJc
UUltldt'il
Kumh'f Hti'f
,it fifhiJ
\ M'Ml'- l!lM»fe
; / :2
] 12-24
1
1
2 0-12
12-24
2 0-32
2*-3§
2 0-12
36-48
2 12-24
24-36
2 12-24
36-48
2 24-34
36-48
3 0-12
12-24
24-36
3 0-12
12-24
36-48
3 0-12
24-36
36-48
3 12-24
24-36
36-48
- pri'Jiding flui*i
ml ski-
ill IV.iJlit
Hu.itlJi'
itiLik*
i.M'F
hi«l.
MuUlt'
Ku,b
MiJJk'
r.i^li
Mid-Ik*
hu'.b
Mill!,.
rfic,h
Midik-
rmh
MiJJtr
r-.igfi
MiJJk'
rns'.h
MiJdU-
Hi^li
MiJJk-
.-llf,lt
Middk-
Fifth
MidJb-
f-IRl. ^
Hit;h
Mi.Uk-
"li^i
MiJdk-
hif,h
MiJJIr
htfji
Mt Idk
Hlj.1,
Mj.Uli
"ti^li
MiJdk-
Hit;!.
MiJJk
rfifth
MiJJIr
Hii'ji
NhJJIc
Hn',!<
".'nh'''
MiJJk
*11S',tl
Mi'dJtc
High
\1idJL'
riii?h
MiJdlr
;si(.',li
\hJJk-
r.mji
4s- ,.» H Mli n jvilL-n .nilnl kiLt-ijr^ l,V ^!)
< :>i-^* 24114
241 ',>,':
rir'\ • i\" • ' ''^ |! i H4 241,' 2.1
I il >.P4
424 %,1 11 i1,,'"^ 24l'.4r
i r1 i';''
2l4 "«\ 0.03 '' T4 2^7'"
i » « f > n
U~ ^^ II 211.2**
« fi. £ u.
?~2 >''C ( 4 .v- i •• 0.02
i "^ i ' r
it, irt- 0.90
i!^ i'l;-'
4,"-4 "H1: <0.01 ''.l^ftl 232.95
"* ^M 'i ' *
20 >'. 4 0.03
1 n4 i'2^i
2^" ti *4 0.12
i ^ (K4
42" 'h;: •: i» "^ tin
,.T-» ltr,2
' -^ r 2 0,30
•;4r> •': '54 0.75
(' 1^ i'¥r
^T 'i." ij I.MH fj,%
135
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE
NOT FOUND IN NRC
(2006)
'Isiblr x "i«i intukt: during?, tin1 titil Vi'.u »H lilr in
pri'Jk tiiii', tluiiRitU :il rrti.vill>Mv a'titia! itui^it*
Cul p; lints Ot
h'VfK ot .ivtni!?;
tluuridi: iritlkx1
during fit 'I \e-.it
of, lil'i1 fir.g/ku, hwi Scti'-iiivitA t<
...Cl v'-^
, .1(4 i 1^1 i
v .»'*> i' T! i
. .vIS ' 17
1.0-
0.9-
0.8- ,-J-''
i;- 11.6- r*..'
^ "-4" /• ""'*'''*
'"-/
ft;"i!li'it\ A*'JUfiK'v
.12 ,!.4fi
1™ jl ^1
.7H \f?.
,"*2 . ,^1
...-•- £f~
-.1-, i.
<,' =Af$tl
i. =i»f?*i
mi Ml M 2 H.3 n.4 it5 n. ti n ~ ii.s in; n
i'i^. j1. C,iiFi'f\itK«iti nt rt'cvKTr ufn-riitinj,1, ^'h-ir-i/lrti'-lir
iKtX'i ^ur'.i-i Idt thmriJi' inl-ikt- Jurint; Iht' iir^t4 VI'.IF--
nl lik',
Of 579 children, 139 (24%) had fluorosis on both maxillary central incisors. Mean fluoride
intake per unit B W ranged from 0.040 to 0.057 mg/kg B W, with higher intake during earlier time
periods and relative stability after 16 months. Inbivariate analyses, fluoride intakes during each
of the first 4 years were individually significantly related to fluorosis on maxillary central
incisors, with the first year most important (PO.0 1), followed by the second (PO.0 1), third
(PO.01) and fourth year (P=0.03). Multivariable logistic regression analyses showed that, after
controlling for the first year, the later years individually were still statistically significant. When
all four time periods were in the model, the first (PO.01) and second years (P=0.04) were still
significant, but the third (P=0.32) and fourth (P=0.82) were not. The study authors concluded
that the first two years were most important to fluorosis development in permanent maxillary
central incisors; however, the study also suggested the importance of other individual years.
Hong, L., S.M. Levy, JJ. Warren, B. Broffitt, and J. Cavanaugh. 2006a. Fluoride intake levels in
relation to fluorosis development in permanent maxillary central incisors and first
molars. Caries Research. 40:494-500.
Russell, A.I. 1961. The differential diagnosis of fluoride and non-fluoride enamel opacities. J
Public Health Dent 21:143-146.
136
January, 2008
-------
PROFILER'S
REMARKS
Initials/D
ate
VAD/12-
29-06
PROFILER'S ESTEM.
NOAEL
PROFILER'S ESTEM.
LOAEL
SUITABILITY FOR
DOSE RESPONSE
MODELING
CRITICAL EFFECTS:
As described in Hong, 2006a, the cohort was 98% Caucasian and 7 1% were from families of
high socioeconomic status; therefore, results are not representative of the general US population.
Incomplete questionnaire data resulted in only 191 subjects available for logistic regression
analyses after four years of life. Fluoride intake data were obtained through serf-administrated
questionnaires by parents without direct verification. The fluoride intake estimates were based on
assessment at 3-4 points during each year and did not account for period variations in intake.
Some potentially important source of fluoride, such as fluoride rinses and gels, were not included
in the intake estimates. The study report did not adequately describe the FRI scoring system, i.e.,
numerical scores for mild, moderate and severe fluorosis. All but 4 of 139 cases were considered
mild fluorosis; therefore, the fluoride intake estimates are only predictive of this gradation, and
are not predictive for severe fluorosis.
Incisors were chosen because they were considered to be aesthetically important teeth; however,
other permanent teeth may be more susceptible to fluorosis.
In children not exhibiting fluorosis of the maxillary central incisors, mean daily fluoride intake
over 48 months was 0.043 mg/kg/day (range 0.038-0.049 mg/kg/day).
In children exhibiting fluorosis of the maxillary central incisors, mean daily fluoride intake over
48 months was 0.053 mg/kg/day (range 0.045-0.062 mg/kg/day).
Not suitable,(_); Poor (_); Medium (X); Strong (_)
Dental fluorosis (maxillary central incisors)
137
January, 2008
-------
Horowitz, H.S., 1989. Fluoride and enamel defects. Adv. Dent Res. 3(2): 143-146.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date:
REMARKS DMO
1/12/07
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
Dental fluorosis
Literature review.
Human and lab animal
NA
NA
NA
NA
NA
NA
NA
NA
Based on currently available information (pre-1989), the author concluded that the
concentration of fluoride in drinking water was the major determinant of the
prevalence and severity of dental fluorosis in a community. The author further notes
that some recent reports suggest that the maturation stages of enamel development are
as important as or even more important than the secretory stages as the time when
fluorosis can be produced.
NA
May be a source of information on early fluorosis studies.
NA
NA
Not suitable (X), Poor ( ), Medium ( ), Strong ( )
Dental fluorosis
138
January, 2008
-------
Ishii, T. and Suckling, G. 1991. The Severity of Dental Fluorosis in Children Exposed to Water
with a High Fluoride Content for Various Periods of Time. J Dent Res. 70(6): 952-956.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS MONITORED:
STATISTICAL METHODS:
Dental fluorosis
Retrospective
Japan/Ikeno District: 86 children, aged 1 1 months to seven years old at the removal of
the high fluoride water (February 1973), that were inhabitants of the Ikeno district of
Japan. The children included in the study were examined at school between 1973 and
1984. Between 1973 and 1981, all permanent residents attending school were
examined annually. The number of subjects varied from year to year. A final
examination of 16 children still at school was completed in 1984. In 1973, 1977, 1978,
1980, and 1984, three photographs were taken of the teeth of each child examined; 41
children had two or more serial photographs available for inspection.
No control population was examined in this study.
Inhabitants of the Ikeno district of Japan were exposed to high levels of fluoride in the
water supply for 12 years, from December 1960 until February 1973. The results
included in this study spanned 11 years, from 1973 to 1984.
The water supply had high levels of fluoride (7.8 ppm) from December 1960 until
February 1973, when it was replaced with low fluoride water (<0.2 ppm. High
fluoride exposure was prolonged (from birth for 7-12 years) for 26 children, restricted
to a shorter time during early tooth development for 38 children and late tooth
development for 22 children.
Other potential sources of fluoride exposure were not included.
Data for measuring the fluoride concentrations were not included in the study report.
Water quality parameters were not reported.
The current study included 86 children in the vulnerable age range (11 months to seven
years old at the removal of the high fluoride water in February 1973), that were
inhabitants of the Ikeno district of Japan. Between 1973 and 1981, all permanent
residents attending school were examined annually. The number of subjects varied
from year to year. A final examination of 16 children still at school was completed in
1984. In 1973, 1977, 1978, 1980, and 1984, three photographs were taken of the teeth
of each child examined. A single examiner carried out all examinations as follows:
Examinations: The teeth were not dried, but were cleaned with gauze if required.
Individual teeth were checked for dental fluorosis using Dean's criteria. The fluorosis
grade of three teeth — the most severely affected of the first permanent molars (molars),
upper central incisors (incisors), and first premolars (premolars) — was assessed for
each child at his/her last examination before leaving primary school. These gradings
were used to determine tooth-type susceptibility to the high fluoride water. Serial sets
of photographs, in combination with the fluorosis score, facilitated assessment of post-
eruptive enamel loss and changes with time for the molars and incisors.
Fluorosis was assessed in all fully erupted permanent teeth following Dean's criteria.
Fluorosis was scaled as normal (N), very mild (Vm), mild (Mi), moderate (Mo) or
severe (S).
No statistical methods or levels of significance were reported in this study.
139
January, 2008
-------
RESULTS:
Fluorosis
Figures 1 and 2 were copied directly from Ishii and Suckling (1991) and summarize
the severity of fluorosis according to Dean's criteria in three tooth types [molar (A),
incisor (B), and premolar (C)]. Figure 1 relates fluorosis to high fluoride water use
from birth and ending at different ages, while Figure 2 relates fluorosis to high fluoride
water use beginning at different ages and ending at eruption or 7 years of age.
* fWSfKMMtaftilOUil
Fit, I — Severity of floorosis m tax torth types assessed icconfing
to Den*! criteria BM| idued to HSC of driakiBg wrtcr «ar»it>mi^ 7,&fpn
P bom Wrtfc (B> feet eading »t diffncnt »gQ. Kombcr of children in cadi
t& group it gfvoL D - N at Q, Q « Vw, Q . M, nx) • »
When high fluoride intake lasted throughout development, mild, moderate, or severe
defects predominated in all three tooth types (Figure 1) with a few exceptions. With
limited exposure, fluorosis resulted when exposure starting at birth continued for
longer than 11 months for the molars, 1 year for the incisors, and 2 years for the
premolars (Figure 1), or when the high fluoride intake began before the age of 3 years
for the molars and 4 years for the incisors and premolars (Figure 2). Moderate or
severe defects predominated among the molars, incisors, and premolars when the high
fluoride intake, starting at birth, continued for more than 2, 3, and 5 years, respectively
(Figure 1). Such severe defects were less evident in the molars and incisors when
exposure started later in tooth development (Figure 2).
140
January, 2008
-------
PERMANENT MQLAH
UPPER CENTRAL
*9«fy) B 1 23456
Number 73331 14
Fij, 2 — Sererity of Hwmsh in ATM too»fa rypc$
P mting K dtfltent agei and
fswUce
- Nor
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/date:
SJG/ 1/17/07
DMO
03/02/07
TAMJE
PREVALENCE CKF KHF-mfPTIW- mUMBL LOSS At 3QWHAL
AUD §3CAMW»TTCi*S OFOfllDKEN EWQSID
TO W*Ttl CONTAINING 7.8 MSB, FLUOSJDE
X ¥
Eipooins feet*
toHJiHF An*
!*,» »;»
C»f» a [Yon) fYi«» tt,4f 11,21 3«**l HJt
J-i'iiiti JHsjIi:
A < toll- f»U 1 1 3 1
B » to6+ KM* 3 3 1 t
c * taSf s.mi: 1 J ? 2
D S to 4+ t,Wl 5441
E ? fej* ?A*.ttU4 6151
r 9 iws* OJA» o Q i §
G 4 felt lAT&U fl Q D Q
* - Ho. dwwhf •» WWWHie io twEnd in* frum im
proEM tt Ac Em nomnnittoit.
The duration of fluoride exposure determined the initial degree of fluorosis, but did not
seem to influence subsequent changes. Pits were restricted to teeth assessed at the first
examination as moderate or severe. No obvious pattern was evident in the subsequent
changes. Teeth graded severe showed variation in the timing and position of any
change. The child exposed throughout tooth development had more severe fluorosis
compared to those exposed only early or late.
PROFILERS NOTE: The profiler agrees that, according to the figures as presented in
Ishii and Suckling (1991), the pattern for development of fluorosis varies among tooth
type and generally is more severe when high fluoride exposure started at birth and
continued for more than 2-4 years and when high fluoride exposure started before 3-4
years of age. However, the statistical significance of such a statement is questionable.
Significant conclusions can not be made.
The information presented here suggests that continuous exposure ending below the
age of 1 1 months or starting above the age of seven years will not result in alteration to
the macroscopic appearance of the permanent teeth (third molars excluded). The upper
age limit, from 4.8 to 1 years, is quite variable. The factors influencing the initial
shape of the pits and the subsequent changes require further investigation. Due to the
small sample size, the conclusions for "at-risk" periods of susceptibility to fluorosis
should be used with caution.
None.
Overall, the paper was difficult to follow and the results were not presented in a
sufficiently clear manner to draw conclusions. The study was not designed for
development of a dose response to fluoride; the emphasis was on monitoring the
severity of dental fluorosis in children exposed to high levels of fluoride in the water
for various periods of time in order to predict critical periods of vulnerability to
fluorosis development. The authors conclude that children were "at-risk" for fluorosis
development if exposure to high fluoride in the water (7.8 ppm) started after 1 1 months
and ended before 7 years of age.
However, there were not enough quantified data to perform any statistical analyses to
support the conclusions set forth in the study. Due to the small sample size (n=l-9 for
each age group with limited exposure), the accuracy of any predicted "at-risk" period
142
January, 2008
-------
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
should be used with caution, as the study authors note. Further, data from photographs
should have been quantified in a manner to include all subjects (41 children had two or
more serial photographs available for inspection), rather than providing only a select
few in the study report.
Other issues to consider in drawing conclusions about the "at-risk" periods of
susceptibility to fluorosis include: Eruption of teeth is governed by circumstances
operating years earlier, with different tooth types developing at different ages. An
accurate retrospective history of fluoride ingestion from all sources (e.g., infant
formula, fluoride supplements, and fluoridated toothpaste) is difficult to obtain;
fluoride/kg body weight is likely to vary.
Study design was not suitable for development of a NOAEL for fluorosis.
Study design was not suitable for development of a LOAEL for fluorosis.
Not suitable (X), Poor ( ), Medium ( ), Strong ( )
The study design was not conducive to provide data for a dose-response. The study
only indicated the prevalence and severity of fluorosis in children exposed to high
levels of fluoride in the water supply over various periods of time. The study did not
address any issues of caries, plaque, or gingivitis. The study is of value, however, in
identifying sensitive age groups.
Prevalence and severity of fluorosis
143
January, 2008
-------
Jackson, R.D., S.A. Kelly, B.P. Katz, J.R. Hull, and G.K. Stookey. 1995. Dental fluorosis and
caries prevalence in communities with different levels of fluoride in the water. J. Public
Health Dentistry 55:79-84.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
Dental fluorosis; dental caries
Prevalence survey
U.S./Indiana/Connersville, Brownsburg, Lowell: Children age 7-14 yr (born between
1978 and 1985). The demographic characteristics of the three communities were
similar, and all three were in the same climatic zone. All the participants had to meet the
following inclusion criteria: be willing to read and sign a letter of consent, and obtain
parental consent; have no factors in their medical history that would contraindicate a
dental examination; meet the criteria for defining lifetime residency; be available during
the examination period; be of proper age at the time of examination; and provide a
residency information indicating that they had used the city water supply or a source
with a comparable fluoride level (±0. 1 ppm) since birth. Lifetime residency was
defined as being born to parents residing in the community and not being absent from
the community for more than two weeks in any one year.
None
7-14 yr, from 1978-1985 until the time of examination in February of 1992.
Fluoride drinking water levels in the three communities were 0.2 ppm (negligible, or
NF), 1.0 ppm (optimal or OFF), and 4.0 ppm (4X OFF). The number of children in each
group are listed in Table 1 copied directly from Jackson et al. (1995):
TAUT F 1
Diiaogf^phJc Dau ui Stuttv Partjdpj^L
AjvRi' gf M. iY.w M P
XT II* '-.•! <^ 1? ^ ••*
'fei 1 '! ~-i-< !"-' -'- l--'i
Drinking water for all three communities was obtained from deep wells. Each
community exhibited a documented water fluoride concentration history for the
preceding 50 years. The optimum fluoride concentration for the geographic region was
1 .0 ppm based on the data presented by Galagan and Vermillion (1 957). History of the
use of fluoride pediatric supplements and commercial products containing fluoride was
evaluated in the study populations by the use of questionnaires.
Methods used for measuring the fluoride concentrations in the drinking water were not
reported. Other water quality parameters (such as calcium levels) were not reported.
Dental fluorosis and caries incidence in children age 7-14 yr (born between 1978 and
1985) from three communities in Indiana having different levels of fluoride in drinking
water (0.2 ppm; 1.0 ppm; and 4.0 ppm) were compared using acceptable methods of
diagnosis (TSIF index and Dean's index for fluorosis, and DMF teeth and DMF surface
scores for caries) and statistical analysis (two way ANOVA).
144
January, 2008
-------
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
Fluorosis was evaluated using the TSIF index and Dean's index (see Section 2 for
descriptions). Only permanent teeth were examined. To distinguish between fluorosis
and non-fluoride opacities, the criteria developed by Russell (1961) were used. Caries
was assessed in terms of DMF teeth and DMF surface scores (Council on Dental
Research and Council on Dental Therapeutics, 1972; see List of Acronyms in report for
definitions).
Scores of Dean's index (minimum of the two highest readings for each child) were
compared across communities and age groups using a two way ANOVA. Statistical
analysis of maximum TSIF scores was performed using the ranks of values, and
comparisons between communities and age groups were evaluated using two way
ANOVA on the ranked values.
As shown in Tables 2 and 3 (copied directly from Jackson et al., 1995), the incidence of
fluorosis increased with increasing fluoride concentration in drinking water.
TABU::
Percent r>l£tr.bi.'Ltia s: Children K'lJcj.-'s Index ikon; and tluorldf Level
y. JDT A? £-\ ,- * Jrdex Score
,?,«•• F. ',i ;i- - : i i
i\f- 2«, i- j, .1 .• ..i/ .•! t: i
O?.~ • !-i •'? .",' VA' f,,J 0 Q
•!\ 3','*- i./ 1 3 ;.•:> Ib.b 1S.J. 32.0 113
TA31.O
?«,f«>r:£ O'lt/lbulior. of Oi.ulfcn bvTSIF Sca.'eii.c Fiuarula l.evft
Kuono- T^rS.-v
, .-.1-1 « ". 1 2 1 -1 n 1 7
\r \J: 8..S i .1: f n r- ; n
ci!- '.i. 5'i.r >j2 •> i :N ..y L i u
'A O?F IT", ~l-i l.~ S ,1 f 2' T A -,' -'-4 ":.;) *,-c
Distribution of TSIF scores by permanent tooth surface (Table 5) showed the same
pattern with higher levels of fluorosis in the 4x OFF community.
T;*BIF=i
jVivSiHJ DJitriiRition af Pi^mi.iiDr.t Tot Hi Sunii;i» £>> i'Bli' Scoje£/.d
»-"ijpr,ce «,«•?!
t .(Iniiiif \t- ol ISrSt'ifi'
k^*c, y'jr^ifrv? P 2 i 1 .* ^ 7
N- -:,^h-' J.= .i - '• l!'i ,) . 000
of»v- -:,i7: ^=,; r-n 2, • .- o o o
4>t^f ?.26f> ?n<; ,\, 5 ;
-------
communities.
TASLF 4
:Ke«if *JH*'tn-br- hJ r7.:-. i-.u ..]..- ^,
. ,-!-. ycarr o: age
NT W i b . I.? 4.1 l
CC'F tf r.fl I ' : T 2 ,
The Community Fluorosis Index (Dean, 1942; see Section 2 for description) was
determined to be 0.15 for the NF community, 0.46 for the OFF community and 2.06 for
the 4x OFF community. Dean indicates that a score of 0.4 or less warrants no public
health concern, while a score of 0.6 begins to constitute a public health problem
warranting further consideration.
Other effects
The mean DMFT score of the OFF group was not significantly different from that of the
other two exposure groups, but the mean DMFT score of the 4X OFF group was
significantly lower than that of the NF group (Table 6, from Jackson et al., 1995). The
mean DMFS scores for both the OFF and 4x OFF groups were significantly lower than
that of the NF group.
TA.iU 6
Me»n D VtFT i.nd .5VtfS 4nw. ;icr CKIJ'Hy -Tv,r»ride level
Hut.'OC. Klt'af! 4> Dif; \t= Ar f^ Ijl^
UNtrS ,SD) lint ,\F
:.54il;:«_ - i; 4.0^^-,- -j, ?
:.y";. t-?1 _ - fiS -i.,1--.T;;.I_ -2V.
Analysis of the DMFT and DMFS scores by age (Table 7 from Jackson et al., 1995)
indicates that only the older children showed significant decreases in these scores at the
higher fluoride levels.
146
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
": AJSUi 7
.\leaii D?wFT j« J I )VlL?S "icwet per CflnJd K- 1- .uwlfce Lcvrf &nti -%««• Group
T rlFl.tl
i'^ <" .">\1F" "i DJ:T DMFP, «? j>,-f
7 ,i V-'-P- ./iv1
%«." r- .-.r , ,.K - iT?;i:is}
v*,1! n'i ».'/'! ^ " * 1 ;>"l.14S1>~j -I?..-
=.\ .JA V> 21"- ft-i - ,7 4,^:2^"' >>:
' 1C \"("s .1: • jl>
^r ivi 'j^!v'''l'i ^.7!l>61^l
'."'«'- -^ 'S Cr,". ~ - Sh .I.fi;i3.'Hi -28J
->. o' " i: r^o^:_ ^37 K ^-f! ,:;<.•.- -SBS
\|T\'-, p,l,i i <•,*,!>, ufO"l..-li.i!r.;cvv,-H(vi
V tli«n5.,', 5.J- te i ji«l n)5' nnr.'H »liljcr V J 'V ,'i
The reported history of fluoride supplement use resulted in an increase in the incidence
of fluorosis (see Table 8 from Jackson etal., 1995), particularly in the NF group, and a
decrease in DMFS scores (in the NF and 4x OFF groups but not in the OFF group, see
Table 9 from Jackson et al., 1995). Based on information provided in the text, the data
in Table 8 pertain to fluoride supplement use during infancy.
TABLE S
Pj.-cet:! cf Subjv«m Afith Maximum "1 Stf Sccre >0 oy XufMirtcti L«.e ot HuorJac
Supp'issrv^ntBiiriiFmafidflevd
r .•'•cc \u^-pp,k.'^-e- - iKfVl S'.ippl^iTt'-is I-'wd
KVI- ?: .\.-/M- 'i \F, XL" '"
\f r r^,"l <••£ ;;, ,v 3,3
t'f'r r~ "'S/*1 ^27 ::v?5 ^:
iXi'i'h 111. >if/L'2 4*3. j 7/fl ,7 .;
•NiJTiS.; rtdr,f-,;,r»T,[-:,',iiii', I,:: .-u.-iuet.
TAJL£«t
Vtej.i I5MPS p»T C InU1 r»r IN.- nf ~,i,.irr,Ju- Supplement* ,-» 7i'; 4 "31, Sj
OFF N't' ** : ',?:"•."=; ~~
. _.!, 2£ .» ^v |2 ^2'
4XOPF No 92 . i,\:. '.*
l'e:- 9 J »>ir
\ i™ 1 " r** -•*« in H4 k* * !X"i -v54*'U '^"im^ t J(i,?u!il a! r'<, 0\
"The ingestion of water containing 1 ppm or less fluoride during the time of tooth
development may result in dental fluorosis, albeit in its milder forms". The milder
forms observed included a maximum Dean's score of 2 associated with white opacities,
but no pitting or staining. Pitting and staining (Dean's score of 3 and 4) were observed
only in the 4x OFF community. The study author observed improper use of fluoride
supplements in communities with optimal and above optimal fluoride levels in drinking
water.
Council on Dental Research and Council on Dental Therapeutics. 1972. Proceeding of
the Conference on the Clinical Testing of Cariostatic Agents. Amer. Dental Assoc.,
Chicago.
147
January, 2008
-------
FOUND IN NRC (2006)
PROFILER'S
REMARKS
DMO/
11/20/06
12/15/06
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
Galagan, DJ. and J.R. Vermillion. 1957. Determining optimum fluoride concentrations.
Public Health Kept. 72:491-493.
Russell, A.L. 1961. The differential diagnosis of fluoride and nonfluoride opacities. J.
Public Health Dent. 21:143-146.
The study authors noted that the 4x OFF group contained 23 children who did not meet
the strict criteria set for this group, some had resided in the study area for only the first 6
years of their life and some had not resided in the study area during their first year of
life; however, the study authors state that the fluorosis and caries scores for these
children were not significantly different from children in the 4xOPF group who were
full time residents. Nevertheless, the study author indicated that these children
thereafter consumed household well water with negligible amounts of fluoride as
defined by the authors' protocol.
The use of fluoride supplements during infancy may have confounded the prevalence of
dental fluorosis in all three study communities.
The study authors did not report on the frequency of use of bottled drinking water in the
study populations.
Severe fluorosis occurred in about 1 1% of the children in the 4x OFF group based on
Dean's index (score of 4) and in about 20% based on TSIF scores (scores of 5-7). No
children in the NF and OFF groups exhibited severe dental fluorosis.
TBD - The NOAEL for severe fluorosis appears to be 1 ppm.
TBD - The LOAEL appears to be between 1 and 4 ppm.
Not suitable (J, Poor (X), Medium (J, Strong (J
The study includes three fluoride exposure levels. Fluoride supplements were
consumed by some of the study participants during infancy (57.9% in the NF group;
19.8% in the OFF group, and 8.9% in the 4x OFF group. This is a potential confounder.
Further consideration should be given to potential statistical evaluation of CFI scores to
estimate the fluoride level where the CFI would be at an acceptable level. According to
Dean (1946, as cited by Jackson et al., 1995), a CFI of 0.6 would be the highest
acceptable score without a public health concern.
Dental fluorosis
148
January, 2008
-------
Khan, A., Moola, M.H., and Cleaton-Jones, P. 2005. Global trends in dental fluorosis from 1980 to
2000: a systematic review. SADJ 60:418-421.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis in children aged 0-19 years
Data were obtained from primary articles published in peer-reviewed journals from January
1, 1980 to December 31, 2000.
Global: only articles with children aged 0-19 years were included in the analysis.
No control population was included.
Only studies with individuals with a life-long residence or those who had lived in the study
area for the first seven years of life were included in the current analysis.
Water fluoride levels were divided into three categories: <0.3 ppm F, >0.3 to <0.7 ppm F,
and>0.7-1.4ppmF.
The exposure assessment consisted solely of reported fluoride concentrations in the water.
It is assumed that community drinking water was measured in each study, but this was not
stated.
The methods for analyzing fluoride in the water of each study included were not described.
Articles were identified in an on-line literature search using PubMed and supplemented by
a hand search using references obtained from articles found in the initial search. For
inclusion each study met the following criteria:
Individuals 0-19 years; both general population and school children were
acceptable, but hospital or clinic samples were not;
Be lifelong residents or had lived in the study area for the first seven years of life;
In an area with water fluoride concentration up to 1.4 ppm;
Have a specified sample size;
Published between the beginning of 1 980 to the end of 2000;
Report fluorosis irrespective of the index used.
From studies meeting the above criteria, the prevalence of fluorosis was pooled and the
trends over time were determined in the three water F concentration categories. Various
indices examined included the Dean Index, the Tooth Surface Index of Fluorosis (TSIF),
Thylstrup-Fejerskov Index (TFI), Fluorosis Risk Index (FRI) and the Developmental
Dental Defects of Enamel Index (DDE) index.
The t-test was used to compare the means of fluorosis prevalence for the fluoridated and
non-fluoridated communities. In statistical analysis, fluorosis indices were used as
continuous variables and the three fluoride concentration groups as categorical variables.
Fluorosis indices were compared with a one-way ANOVA. For categorical variables a
Chi-square test was done with Cramer's V (to find significant differences in the distribution
of scores of the different indices). A Bonferroni test was used to determine differences
between the means of the proportions of the prevalence.
149
January, 2008
-------
Fluorosis
Literature evaluation
Dose response
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date
REMARKS CSW1/4/2007
PROFILER'S ESTEM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
The mean percent of fluorosis prevalence increased with increasing fluoride concentration
(Table II). The prevalence was significantly greater in the >0.3 to <0.7 ppm (p=0.041) and
X).7-<1 .4 ppm (p=0.020) groups compared with the 0-O.3 ppm group.
Graphical representation of fluorosis prevalence vs year of publication showed slight
upward trends over time at all three fluoride concentration levels.
Table II: PareentasB pravalanoo rates of fluorosls by fluoride concentration category |
Fluoride Publications
concentration n
Ppm
0 to <0.3 49
>0,3 to <0.7 9
>Q.7to<1.4 37
Mean SD
%
16.7 178
27 4 32 ?
32 2 23 5
Minimum Maxsmum
% %
0 78.Q I
2.4 93,7
6-6 87,6
A total of 55 publications met the inclusion criteria for this review. Five fluorosis indices
were used in the various publications such that too few studies per index were available to
assess severity changes over time. Thus, the review concentrated only on fluorosis
prevalence since little difference existed in the criteria between indices for the decision of
fluorosis vs no fluorosis.
Increasing fluoride water concentrations were associated with higher prevalence of
fluorosis.
PROFILER'S NOTE: No attempt was made by the authors to estimate fluoride intake
from drinking water consumption; thus, a fluoride dose could not be determined.
This review confirms observation of an increase in fluorosis prevalence among fluoridated
and non-fluoridated communities, although the trend is not statistically significant. Wide
variation in fluorosis prevalence implies exposure via non-water sources such as
fluoridated salt, beverages, food, toothpaste, dental rinses, etc. Total exposures "must" be
taken into account for control of fluorosis. Residence at high altitude may also be a factor
in developing dental fluorosis.
Limitations of the review include differences in study quality as well as in the diagnostic
criteria employed among and between studies.
none
This study was a review and analysis of data in published literature. The authors used
specific inclusion criteria to choose studies. However, the quality of the individual
publications was not considered.
Doses could not be reconstructed from this review.
Could not be determined.
Could not be determined.
Not suitable ( ), Poor (X), Medium ( ), Strong ( )
A positive correlation was found between increasing water fluoride levels and increasing
fluorosis prevalence; however, a dose was not estimated. The severity of fluorosis was also
150
January, 2008
-------
CRITICAL EFFECT(S):
not assessed.
Dental fluorosis
151
January, 2008
-------
Kliniek, J., H. Prinz, E. Hellwig and G. Ahrens. 1985. Effect of a preventative program based on
professional toothcleaning and fluoride application on caries and gingivitis. Community Dent
Oral Epidemiol. 13: 295-298.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL
POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL
METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
Dental caries, plaque accumulation and gingivitis (no fluorosis observed)
Case-control study
104 (50 boys and 54 girls) schoolchildren, ages 12-14 years old, that were part of the resident
population from Marburg, Germany. Most were from families in the higher socioeconomic status.
1 17 (52 boys and 65 girls) schoolchildren, ages 12-14 years old, that were part of the resident
population from Marburg, Germany. Most also were from higher socioeconomic families.
The study occurred over a period of 2 years with the first examination taking place in 198 1 when
participants were between 12-13 years old.
The two exposure groups were defined by the level of professional tooth-cleaning and
prophylactic dental fluoride treatment received. The test group received a prescribed regime of
prophylactic dental care and the control group received none. Children were also exposed to the
drinking water from the town of Marburg, Germany which had a negligible (< 0.2 ppm F-)
amount of fluoride, but the amount of water consumption was not quantified.
All children were examined for dental plaque, gingivitis and caries present over a 2 year period
with the test group receiving prophylactic care during that time (See details below). Radiographs
were taken before the start of the study and approximately 2 1A months after the test group
received their last prophylactic treatment.
Data on how fluoride concentrations in the water were measured were not included in the study
report. The fluoride concentration in the varnish applied to the test group was also not included.
In the study, 104 (50 boys and 54 girls) schoolchildren, ages 12-14 years old, made up the study
population and 1 17 (52 boys and 65 girls) schoolchildren, also ages 12-14 years old, made up the
control population. All children were part of the resident population of Marburg, Germany. The
two groups underwent different protocols. The study population underwent dental examinations
but also received prophylactic treatment; the control population only received dental
examinations. Both procedures were as follows:
Examinations: Prior to the start of the program and about 2 l/i months after the last prophylactic
session for the test group, both the control and test groups were examined clinically and
radiographically. Examinations consisted of an assessment of plaque, gingivitis and caries
performed by a single examiner. Assessments were performed using an artificial light source,
dental mirrors and dental probes. Radiographs taken were two posterior bitewings and were
evaluated by two dentists that were blinded to which group the children belonged.
Prophylactic treatment: The test group children visited the same oral hygienist 4 times in the first
6 weeks and then visited the same oral hygienist 5 times/year to receive professional oral
prophylactic treatment and instructions. Each session included staining for dental plaque,
demonstration of tooth-brushing and flossing and professional tooth-cleaning. The abrasive paste
used in the cleaning contained no fluoride. Also, a fluoride varnish (Duraphat) was applied twice
a year. At the end of each visit, the children were provided with large quantities of toothpaste
(each containing NaF), toothbrushes and dental floss. The control children received no
prophylactic treatment.
Prior to the start of the program and about 2 1A months after the last prophylactic session for the
test group, both the control and test groups were examined clinically and radiographically.
Evidence of dental plaque (PI I) used the criteria and indices of Silness and Loe (1964) with
152
January, 2008
-------
STATISTICAL
METHODS:
RESULTS:
Caries data
Plaque and gingivitis
clinical and radiographic caries recorded according to the indices of Koch (1967). The DPS index
(number of new carious and filled surfaces) was used to describe the caries incidence. Gingivitis
(GI) was evaluated as proposed by Ramfjord (1959) and Loe and Silness (1963).
During the prophylactic treatment, each session included staining for dental plaque,
demonstration of tooth-brushing and flossing and professional tooth-cleaning as described by
Axelsson and Lindhe (1974).
Statistical significance between the test and control groups for plaque, gingivitis and DPS were
analyzed with the Mann-Whitney U-test. The Wilcoxon test was also used for determining the
significance of differences in each group between the initial and final examination.
Tables 1 and 2 are copied directly from Klimek et al. (1985) and demonstrate the pre- and post-
experimental incidence of caries in the test and control group. Table 1 indicates no statistical
significant difference between the control and test group in the incidence of caries prior to the
start of the prophylactic treatment. Table 2 shows a statistically significant (p value not given)
increase in the incidence of caries in the control group after 2 years compared to the test group on
all surfaces of the tooth (occlusal, proximal and buccal-lingual). Data shows the control group
children developed nearly twice as many carious and filled tooth surfaces when compared to the
test group.
*^ !ik- Pic i *|v i t wsi ,1 u1 U"* dti 1 u'j te^ tit LO( no' t! Itae %k*isUi uiuli^it-^ itim •* '&f't aj t ik'Ui is
*>> t "UsM" i(fi k. ,sr i?i imup Sti — P l
Sn f«tOi^ *„*•>!> MM n * + bD M«usu) < ^fljr.u re*
If l,il -i-> i.l ,i' i ,• 1 *;•> -i~ li frs T [ 1 ".,1 * i si
N i On. .i,> -Iduro fiininit 2 ^ i •» . , su :v -2 S \S
su ,.- a>ii) iiui h tui . ,J1 «i J."'% 5 Ti ^ , J-6n •,« \S
Rst ^[XMD l>f- P^ ronlrnj jm j^ ;?, !'?( |A»vci af
>j^'.ttcs i^*O N\cj" « ^_SD \t !tdf agaifisHice
InM ]<• nlMIrl.si. i;5)±^" > .' Z1!^!^! if 1 NS
\n i'ii.» Xi u-rf' l""±i- i SW .' '2 t <» I J7 ***
id u>i> -Hied n*i H i t~'-±.^ (i i4 :« j * iv ***
sallu.iDH'j) .' t\<) IT] (1'- ! I2"1 J1-, 14 «3S **
•>ti -",-2-> I •r Mcars
?r U'sd fekmUart. tJe^iauun fSL>f
!Jfc»M> tnc cv n itijtian I s>lkm~u~s t:\aniinat t* i
1 > "W 1 t , ! VH i 1
\Ln Ma\ i_iSD V,., M,,\, \iSI> P
Tcs; sjroap PI I JS 2 b i 5 t t» 4 >> U 14 US_,l.l --«!>!»•
i« ;cJ} t,ii o. 2; i.;^_(i.4 n.it 15 i»5_,-Ji < a.oa:
Cona-o' grol,p PI I U4 2 *> 1 ^ +-(» ^ t» 7 ? -, ^iO". NS
i». ,17)" GI U.O 1 N I'.'' If'.-J 11,3 ! ^ 1 i-r-O,-i '- "i 35
153
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE
NOT FOUND IN NRC
(2006)
PROFILER'S DFG/11-06
REMARKS and
12/15/06
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM.
LOEL/ LOAEL
POTENTIAL
SUITABILITY FOR
DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
PROFILER'S NOTE: The profiler agrees that the test group had decreased plaque and gingivitis.
Klimek et al. (1985) concluded that a preventative program based on five prophylactic sessions
per year and the additional application of fluoride varnish twice a year was remarkably effective
in reducing gingival inflammation and the development of new carious lesions in schoolchildren.
The study showed a caries reduction of 46% and a reduction of PI I and GI mean index values of
about 60%.
Axelsson, P. and J. Lindhe. 1974. The effect of a preventative programme on dental plaque,
gingivitis and caries in schoolchildren. Results after one and two years. J. Clin. Periodontal, 1:
126-38.
Koch, G. 1 967. Effect of sodium fluoride in dentrifice and mouthwash on incidence of dental
caries in school-children. Thesis. Odontol. Revy; Supple. 12.
Loe, H. and J. Silness. 1963. Periodontal disease in pregnancy. Acta Odontol Scand, 2:533-51.
Ramfjord, SP. 1959. Indices for prevalence and incidence of periodontal disease. J. Periodontal,
30:51-9.
Silness, J andH. Loe. 1964. Periodontal disease in pregnancy, II. Correlation between oral
hygiene and periodontal condition. Acta Odontol Scand, 22:121-35.
The study was well-conducted and had adequate study design. However, the study was not
designed for development of a dose response to fluoride as the emphasis was on a total
prophylactic program (brushing, dental care, flossing) as well as direct application of a fluoride
varnish twice a year. Based on the study design, the observed decrease in incidence of dental
caries, plaque and gingivitis cannot be attributed to any one cause.
Study design was not suitable for development of a NO AEL or LOAEL for fluorosis.
Study design was not suitable for development of a NO AEL or LOAEL for fluorosis.
Not suitable (X), Poor ( ), Medium ( ), Strong ( )
While the study was well-conducted, the study design was not conducive to provide data for a
dose-response. The study only indicated that a strong prophylactic program, including the use of
fluoride treatments twice a year, does help decrease the incidence of caries, plaque and gingivitis.
The study did not address any issues of dental or skeletal fluorosis.
Incidence of dental caries, plaque and gingivitis
154
January, 2008
-------
Kaur et al. 1987. Changing trends of dental caries and enamel mottling after change of fluoride
content in drinking water in endemic fluoride belt. J. Indian Soc. Pedo. Prev. Dent, pp.37-44.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Caries
Dental caries and dental fluorosis
Cross-sectional study of dental fluorosis and dental caries and fluoride levels in drinking
water
India/Punjab: 988 children (sex not specified, ages 6-16 yrs old) from three villages. No
other information concerning the study population was provided.
None
From birth to 6-16 yrs old.
3 groups: Group 1 (town of Naruna), previously 2 ppm fluoride in drinking water, changed 9
yrs later to 0.19 ppm fluoride; Group 2 (town of Deon) previously 1.8 ppm fluoride,
changed 11 yrs later to 0.19 ppm fluoride; Group 3 (town of Bibiwala) previously 0.9 ppm
fluoride, changed 8 yrs later to 0. 19 ppm fluoride
Drinking water was the only exposure route evaluated.
Orion Fluoride Selectrode was used to analyze for the fluoride in drinking water
Dental caries and fluorosis were investigated in Indian children, ages 6-16 years old, in three
villages: two where the initial fluoride levels were 2.0 and 1.8 ppm, and a third where the
fluoride level was 0.9 ppm, but a change in the water supply brought all three areas to 0. 19
ppm fluoride. Each child was evaluated for dental caries using Moller's Index (Moore,
1966) and the degree of dental fluorosis using Dean's Index of Fluorosis (see Section 2).
The prevalence of severity of the dental caries and dental fluorosis was assessed, broken
down by the age of the children, and this was correlated with the years that the water supply
fluoride levels changed in each area.
Dental caries were measured using Moller's Index (Moore, 1966) and dental fluorosis using
Dean's Index of Fluorosis (see Section 2)
Not stated
See Table 5 for the prevalence and severity of dental caries in Village 1 - Naruna (previously
2 ppm fluoride, present 0. 19 ppm fluoride), Table 6 for Village 2 - Deon (previously 1.8
ppm fluoride, present 0.19 ppm fluoride), Table 7 for Village 3 - Bibiwala (previously 0.9
ppm fluoride, present 0.19 ppm fluoride) (all tables copied directly from Kaur et al., 1987).
155
January, 2008
-------
TABLES
POINT PiEV ALENCE, DMFT AND DMFS OF FUST PERMANENT MOLAR IN VILLAGE
NARUANA
Age Total No. Point Prevalence DMFT DMFT
(Yrs) of Teeth No. % Mean S.D, Mean S.D,
Examined
6 "2 7? 68.7 2.75 1.20 3.82 1.72
7 88 51 57,9 2.32 im 2.63 1.18
8 U2 6? 59.8 2.39 1,1? 2.88 1.56
9 136 72 52,9 2.12 1.58 3.19 2.69
W H6 53 45.7 1.83 1.38 2.70 2.31
11 92 42 45.6 1.83 0.71 2.04 0,77
12 *68 73 43.5 1.71 1.27 2.46 1,»
13 136 60 44.1 1.76 1,61 2.13 2.10
14 132 56 42.4 1.70 1.48 2.37 2,15
15 180 73 «,5 1,62 1.05 1.92 1.31
16 MO 59 42,1 1.69 1.11 2.34 1,90
TABLE 6
POINT P1EVALENCE, DMFT AND DMFS OF PERMANENT MOLAR IN VILLAGE
DEGN
Age Total No, Point Prevalence DMFT DMFT
(Yrs) of Teeth No, % Mean S.D. Mean S.D.
Examined
6 120 64 53.33 2.13 0.4S 2.50 0.54
7 132 61 46.21 1,84 0.82 2.05 0.89
8 112 56 50.00 2.00 01.05 2.27 1.3?
9 life 51 44.30 1,76 0.82 2.05 1,03
10 116 50 43.40 1.72 1.10 2,01 1,48
H 84 36 42.40 1.71 0.76 253 1.29
12 1% 73 37,20 1.48 1.13 2.78 3,01
13 168 56 33.30 1,32 1.33 1,97 2,20
11 120 52 43.30 1.72 1,46 2.46 2,50
IS 88 40 45.50- 0.56 1.28 1.39 3.17
16 60 24 40.00 1.60 0,94 2,08 1.36
F
n
11 " ""™ — •• ", —
TABLE f
POINT PREVALENCE,. DMFT AND DMFS OF HRST PEiMANENT MOLAR IN VILLAGE
BIBfWALA
Age Total No. Point Prevalence "DMRp" ' ~ 7^T~~ —
{Yrs) of Teeth No * u c- DMFT
E».m«d % Mean S-D- "««» S.D,
7 M !? 69'° 2M « «6 0 95
; i • :" - s s :::
= : i ;'i !i = =
5 = ; S S S = =
ULi_lJLljJLjL
ROFILER'S NOTE: It appears that the last two columns in Tables 5-7 refer to DMFS anc
otDMFT.
[
156
January, 2008
-------
Dental fluorosis
S
fc
ee Table 2 for the prevalence and severity of dental fluorosis in Village 1 - Naruna, Table 3
)r Village 2 - Deon; and Table 4 for Village 3 - Bibiwala:
JAMS. 2
PiEVALENCE AND SEVERITY OF ENAMEL MOTTLING IN FKST PEtMANENT
MOLAR IN VILLAGE NARUANA
Age Total No, Teeth Mottled GRADES OF MQTTI ING
(Yrs) of Teeth 0 I 2 3 44
Examined No, % No. % No, % No, % No, % No. %
6 H2 0 0.0 112 100.00 0 0,00 § 0.00 0 0,00 0 • 0.00
7 88 0 0.0 88 1SO.OO 0 0,00 0 0.00 0 0.00 0 0.00
8 112 0 0.0 112 100.00 0 0.00 0 0.00 0 0.00 0 0,00
9 136 130 95.59 6 4.41 2 U7 128 94.11 0 0,00 0 0.00
M 116 103 88.79 13 11.21 § 0.00 103 88.79 0 0.00 0 0.00
11 92 82 89.13 10 10.86 0 0.00 82 89,13
12 168 144 85.71 24 14,29 2 1.19 142 84.52 0 0.00 0 0.00
13 136 134 98.53 2 1.47 0 0.00 134 98.53 0 0.00 0 0.00
M 132 123 93.18 9 6.82 0 0.00 119 90,15 4 3,03 0 0.00
15 180 180 100.00 0 0.00 0 0.00 176 97.7S 4 2.22 0 0.00
16 140 140 100.00 0 0.00 0 0.00 140 100.00 0 0.00 0 0.00
TABLE 3
PREVALENCE AND SEVERITY OF ENAMEL MOTTLING IN FUST PERMANENT
MOLAR IN VILLAGE DEON
% 5SS- Te€*MOttef 0 GRADES-HTMOFTUNG— —
EnM No. % No. % No, % No, % No 3 % No«%
* J20 0 0.0 120 100.00 0 0.00 0^^ 0^~7~W
J 2 0 0.0 132 1.00.00 0 0.00 Q 0,00 0 aw Q
8 « « 5.36 106 94.64 6 5.36 0 0.00 0 (,„ 0 000
10 6 ! L "6 mm ° °-w ° OM ° ao° ° oS
? "6 ° OJ m mM o o-w o 0.00 o ooo o Jon
m m «,° 1 ^ ° °m n l4'29 9 lftn ! "
,f 1% m 95'g 8 4.08 1 0.5! 186 94.9 1 051 0 oon
£ m moo o 0.00 o 0.00 m moo S ! ow
5 « M S2 ° °'W ° «» m ^ ° ^ 0 «
« 88 88 100.00 0 0.00 0 0,00 84 95.45 4 460 0 OW
_» « 60 ».» 0 0,00 o 0.00 « 100.00 0 lm 0 a»
, J
157
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date
REMARKS SBG
3/28/2007
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
TABLE 4 ~~~ ' ~~ ~
PtEVALENCE AND SEVEBTY OF ENAMEL MOTTLING IN FUST PERMANENT
MOLA1 IN VILLAGE WBIWALA ™™»^«
ffi> 2S5T Te€thM°"W o "«SADis-5F-MHTTLlT5G-— - ~
— ~-^!l^-^^_^ * *>• * No. 2 % Ncx 3 % No 4\
7 « ! 1° 10° mm ° °M ° °'W ° 0-00 ' o 000
I " ° °*° m mm * 0-00 0 0.00 0 0,oo o 000
98 « • jo «» mm o o» o ».«, „ O.oo o J2
10 m I nn I2 m°° ° °-°° 8 mo° ° °'°° o ».oo
,, 0 0.0 88 100.00 0 0.00 0 000 0 000 0 om
£ ? ?« * 77J8 ° om " 22- « ™ » "
i, 1 6M m 93JS ° 0-00 7 6.25 0 000 0 000
S 2 ^ 79 ^ 0 «"» « 24.« 0 0.00 ow
« "6 ° °'° m 1•» o 0,00 o 0.00 o ooo
! J °ft° K mm ° '» o 0,00 o 0,00 o o«
92 ° 0'° 92 'o0-00 ° 0.00 o o.oo o o.oo o nm
]
Regarding dental fluorosis, in Village 1 (where the fluoride level was lowered from 2 ppm to
0.19 ppm), grade 3 dental mottling was observed in only 3% of 14 year old children and
2.2% of 15 year old children. In Village 2 (where the fluoride level was lowered from 1.8
ppm to 0. 19 ppm), a similar picture was seen, however grade 1 mottling was seen in more
children than in village 1 . A low prevalence of grade 3 mottling was seen only at 1 1 , 12, and
15 years of age. In village 3 (where the fluoride level was lowered from 0.9 ppm to 0.19
ppm), a low prevalence of mottling was observed in 9, 1 1, and 13 year old children. Grade 3
mottling was not seen in any child.
Regarding dental caries, in Village 1, the percentage of first molars with caries in 9 year old
children was 52.91, in children younger than 9 years old the percentage of carious molars
ranged from 57.9 to 68.7, in children older than 9 years old, the percentage was 40.545.7. In
Village 2, a similar picture was observed, with the 6-10 year old children showing a higher
prevalence than the 12-16 year olds (with the water supply being changed at 1 1 years). In
Village 3, a trend of sudden increase in decayed, missing, and filled (DMF) teeth and
prevalence of carious molars after the year of the change of the water supply was also
observed.
Moore, I.J. 1966. Clinical criteria for the diagnosis of the incipient caries lesion. Advanc.
Fluori. Res. 4: 57-72.
This study is limited in its usefulness because no statistical analyses were carried out.
Although confounding factors which may have contributed to total fluoride intake were not
addressed, the study does suggest that for the study population a NOAEL for severe
fluorosis corresponds to about 2 ppm fluoride in drinking water.
Based on the results in Tables 2, 3 , and 4, a NOAEL of 2 ppm for severe dental fluorosis
(Grades 4-6) can be identified, because none of the children in Village 1 (even the older
children who were drinking water at 2 ppm fluoride when their teeth erupted) showed this
degree of dental fluorosis.
Based on the results in Tables 2, 3, and 4, a LOAEL for severe dental fluorosis cannot be
identified from this study, because none of the children in any of the 3 villages showed
severe dental fluorosis.
158
January, 2008
-------
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
Not suitable ( ), Poor ( x), Medium ( ), Strong ( )
This study seems to be poorly suited for dose-response modelling because even though a
NOAEL was identified for dental fluorosis, a LOAEL was not identified, and no statistics
were carried out on the data.
Dental caries, dental fluorosis
159
January, 2008
-------
Kumar, J.V. and Swango P.A. 1999. Fluoride exposure and dental fluorosis in Newburgh and
Kingston, New York: policy implications. Community Dent. Oral Epidemiol. 27: 171-180.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
Dental fluorosis
Cross-sectional
U.S./New York/Newburgh City (fluoridated water supply): 7-14 yr-old school children
(lifelong residents); 459 (47.9% male; 51.6% African-American) surveyed in 1986; 847
(49.0% male; 41.1% African- American) surveyed in 1995.
U.S./New York/Newburgh Town (fluoridated water supply): 7-14 yr-old school children
(lifelong residents); 289 (58.8% male; 8.3% African-American) surveyed in 1986; 289
(49.1% male, 10.4% African- American) surveyed in 1995.
U.S./New York/New Windsor (non-fluoridated water supply): 7-14 yr-old old school
children (lifelong residents):-; 134 children (52.2% male, 5.2% African-American)
surveyed in 1986; 237 (41.8% male, 6.8% African-American) surveyed in 1995.
U.S./New York/ Kingston (non-fluoridated water supply): 7-14 yr-old school children
(lifelong residents); 425 children (50.3% male; 16.0% African- American) surveyed in
1986; 646 (50.8% male, 19.2% African-American) surveyed in 1995.
U.S./New York/Ulster (non-fluoridated water supply): 7-14 yr-old school children
(lifelong residents), 174 children (50.0% male; 4.6% African-American) surveyed in
1995.
PROFILER'S NOTE: Although the children in New Windsor, Kingston and Ulster did
not have exposure to fluoridated water, some had exposure from fluoride supplements
(tablets) and dentifrices with fluoride (see Table 1 under Exposure Groups).
Children from Newburgh City were exposed throughout their lifetime, except during a
three year period from 1978 to 1981 (affecting children from the 1986 survey as follows:
7 year-olds from 0-2 years of age; 8 yr-olds from 0-3 years of age; 9 yr-olds from 1-4
years of age, etc., ranging up to 14 yr-olds affected from 6-9 years of age).
Children from Newburgh Town surveyed in 1986 were exposed to fluoridation for 2
years. Those surveyed in 1995 were exposed beginning from birth up to age 3 for a
maximum of 1 1 years.
Children from New Windsor, Kingston, and Ulster were not exposed to fluoride in the
drinking water supply.
Table 1 was copied directly from Kumar and Swango (1999) and summarizes the percent
distribution of children in each study community according to age group, gender, race,
and fluoride exposure. Males and females generally were equally represented in all areas,
although there were differences in gender between 1986 and 1995 in Newburgh Town
and New Windsor. In both 1986 and 1995, there were proportionally more African-
American children in Newburgh City compared to other areas. The major (80-95%) race
represented in each community was White or 'other', with 5 to 20% African- American;
the exception was Newburgh City where African- Americans made up 52% and 41% in
1986 and 1995, respectively
160
January, 2008
-------
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN:
Table I, Percentage distribution of children in vudv rommiiratil*. a .tort 11% «. age j?n up RCiKk-r ra«.i *n,I fiuo-idi- f>pw,un;
Vwbu*ghClty Nk'wbmjth lown New WtnJ«>r Kingvlon UKhir*
W,* W5 1-S6 W 1«86 "OS IW, I«S !<»S
Jtomuten stetuv 1= f M- F %!F NF XF MF NT
NumVr «•' s*™ ;(W 2!W !34 2T* 42*' M6 '""*
Age group 7-lf) SI* «* t7» 54- S5S -ill 497 W3 Ml t
'* ** F ;] ,4 4S-- 41! 4 «I 451 ttS IS » 503 40' fc' S
5,,, Mak. 47'> 4"d •«« «t K2 41 S 501 W8 H><>
Ivtpilt 521 SHI 4)2 SO* 47* 58! ->97 492 50 U
RM WhK.'w ,nd othe i 4M 5S « •>! " «",!» «J Mil M« 80S IS 4
Afntun-Aiwrica JH. 417 * 3 1'H 53 h» l<- (I i»2 46
Enposure n«ondjtion4ul d MM 4V J 5K ">
and/or early t ythmg
Hu-iridit.ononl 4">" 508 -4)1
(••«ivtlrsh«nS!U1 "*S - - 41 ; %'." 433 .146 UJ) 348 V-. 1
T,U,,, - 100 76' 3"" MR U* W5 1^5
fv' } fh ' •abtl^ i - - *8 ^ 33 * 32 S V ' 3*> 1 3i S
1 FnfluonrtaWd Ml- npnfl.Kindatcd
2 Children cxammtd m Nc«i»uf)»h City in l'»W> fad an ntiTMp>n r in fli»iri<•!»! tel and/c-r t>m»l-in(; c*t»f on
^ D»I* fcr 198!) mil Biiilibk-
Some children in all the surveyed communities also were exposed to fluoride from other
sources, including fluoride tablet supplements and early brushing (before age 2 years)
with fluoridated toothpaste.
In the non-fluoridated communities 9see Table 1), approximately one third (29-35%) of
the children were not exposed to fluoride from any source considered (water, tablet, early
brushing); approximately another third (3543%) were exposed only by early brushing;
the remainder were exposed via tablet only (4-15%) or tablet plus early brushing (17-
21%).
Data for measuring the fluoride concentrations were not included in the study report.
Water quality parameters were not measured. The fluoride concentration in supplement
tablets or toothpaste was not included.
The purpose of the study was to determine whether the risk imposed by fluoride exposure
has changed during the time period from 1986 to 1995, and to determine the effect of
water fluoridation and other known fluoride sources (fluoride levels not reported) on
dental fluorosis in participating children. Data were obtained from two cross-sectional
surveys, conducted in 1986 and 1995 during the school years in the Newburgh and
Kingston school districts. Analysis was limited to 3500, 7-14 year old lifelong residents
of Newburgh City (fluoridated at 1±0.2 mg/L since 1945, with a 3 year interruption from
1978 to 1981), Newburgh Town (fluoridated since 1984 — level not reported), and New
Windsor, Kingston, or Ulster (all non-fluoridated).
The two surveys were similar in design except for the number of children studied
(n=1307 in 1986; n=2193 in 1995). Dental fluorosis and caries were recorded using
Dean's criteria and the Decayed, Missing and Filled Tooth Surfaces (DMFS) index.
Examination details were not reported, including the examiners, location where
examinations occurred, lighting conditions, or equipment used. Examiners were blind to
residential and fluoridation status of the children. Fluoride exposure data was collected
by questionnaire. The questionnaire design differed between studies as follows:
1) The responses in the 1986 survey were primarily open-ended for the two
questions related to the use of fluoride toothpaste and fluoride tablets. In the
1995 survey, categories for these two items were created.
2) Questions related to school lunch program participation, education of the head of
the household, and breast-feeding were not included in the 1986 survey.
3) The response rate for participation was lower in 1995 (38%) than in 1986 (58%).
161
January, 2008
-------
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
4) Children from the town of Ulster could not be identified from the 1 986 survey
because a question regarding their place of residence was not included.
5) Different examiners were used in each survey, although they were trained by the
same dentist.
PROFILER'S NOTE: Dean's Index of Fluorosis is described in Section 2 of the report).
Dental fluorosis was measured using Dean's classification at the subject level. Fluorosis
was scaled as none, questionable, very mild, mild, moderate or severe. Dental caries was
recorded using the DMFS index (number of decayed, missing and filled permanent tooth
surfaces) and visual-tactile examination.
Regression procedures were used to estimate the effect of fluoridation, fluoride
supplements, and early brushing on dental fluorosis. Analysis involved comparison of
frequency distributions of Dean's fluorosis categories to examine changes over time,
including a "ridit" analysis to examine the changes in severity. The proportion of
children exposed to known fluoride sources was calculated for each category: a)
fluoridation alone; b) fluoridation plus tablet and/or early brushing; c) tablet supplement
alone; d) early tooth brushing alone; e) tablet supplement plus early brushing; f) none of
the above exposures. Adjusted odds ratios and 95% confidence intervals were calculated
for the variables associated with dental fluorosis in the bivariate analysis (p<0. 1). This
model included race and the categories for fluoride exposure variables. The fluoride
exposure variable was introduced as an indicator variable and the reference group
consisted of children from the non-fluoridated areas without known fluoride exposure.
Logistic regression procedures were used to estimate the association between factors and
dental fluorosis separately for the two surveys, and then were used to fit a generalized
"logit" model for examining the effect of year, race and fluoride exposure and their
interactions.
Two separate regression models were constructed to compare categories of fluorosis: one
to compare the questionable category with the normal, and another to compare the very
mild to severe category (combined) with the normal. The examination of two-way and
three-way interaction terms requires larger sample sizes, so exposure categories were
combined as follows: a) fluoridation alone or fluoridation plus tablet and/or early
brushing; b) tablet supplement alone or early tooth brushing alone or tablet supplement
plus early brushing; c) reference group with none of the above exposures. The purpose of
this analysis was to determine whether there were changes in the risk associated with
fluoride exposure between the two surveys for African-American children and children of
other racial groups.
Analyses of crude, covariate-adjusted and ranked DMFS scores were performed to
determine the relationship between caries and dental fluorosis. Four categories of
fluorosis (normal, questionable, very mild, and mild to severe) were created for this
analysis. Other variables included in the model were age, poverty status (based on
participation in free-lunch program) in fluoridated and nonfluoridated areas, presence or
absence of sealants, and education of the head of the household. The dependent variable
(DMFS) was converted to a rank and the tests of significance were based on the rank of
analysis of covariance.
Tables were copied directly from Kumar and Swango (1999). Table 2 shows the
distribution of dental fluorosis according to Dean's index by place of residence and year
of examination. The highest prevalence of the very mild to severe categories was
observed in Newburgh City in 1995 (18.6% combined, vs. 14.8% in Newburgh Town,
14.4% in New Windsor, 11.1% in Kingston, and 14.4% in Ulster). Between-survey
comparisons show that neither the prevalence nor the severity of dental fluorosis
increased after Newburgh Town was fluoridated. The case was the same for non-
fluoridated areas. Changes were evident in Newburgh City (fluoridated), where a ridit
analysis showed that the odds were 4 to 3 that a child examined in 1995 would have at
162
January, 2008
-------
least questionable fluorosis, compared with a similar child in 1986.
Table 2 Dental fluorosfe prevalence in percent .crardtag to Dean's classification by place of residence ad year of exi-mination
Cfean's index
Wa<.e %„ Status Number Normal Questionable Very mild Mild Moderate Severe^
„, >, uf~. iaa F 4"» 784 13.7 4-8 2,2 0.9
Newburgh City iwo r ***>? iu^*
J9J51 F 847 6Z-* 18-3 *2-8 w *Jl<*
Ncwburgh Town 1986 NF 2® 73.0 13,1 8,7 4.2 1-0
New Windsor V» NF 134 76-1 9-7 « « «
1995 NF 237 7S.5 10,1 8-9 5,1 0.4
Kini«« 1*4 NF 425 81.4 11.3 4.3 2.1 0.7
IW5 NF M6 | 81-4 M '-' J-' U'J
Ulster 1986 NF - -^ 98 2~9 11 06
' Ridit for Hsb group «v« 0.58 relative to 0.50 tor the 1W survey (statistically sipiificanl, F<0.05). AH ofcr betwcen-survey
comparisons vieldtd ridit values o< less than 051 . .
' The pimalence of very mild to severe dental fluorosis among children bom after the ,mplemerrtat.on of fluoridation ir.
Ncwburgh Town was 14.7 (37/252).
Table 3 summarizes the adjusted odds ratios associated with fluoride exposure and race
by year of study. Children using fluoride tablets and early brushing had the highest odds
ratios for very mild to severe fluorosis in both 1986 (OR: 5.0; CI: 2.5, 10.2) and 1995
(OR: 4.0; CI: 2.4, 6.9). Elevated odds ratios were observed for all the fluoride exposure
variables in both years; however, exposure to fluoridation alone in 1986 was not
statistically significant. African-American children studied in 1995 were at higher risk
(OR: 2.3; CI: 1.8, 3.0) for dental fluorosis than children of other racial groups. While
elevated odds ratios were observed for questionable fluorosis in 1995 (range: 1.6 to 4.4),
they were not statistically significant in 1986 (range: 1.0 to 1.5).
JMf J (Md-i tatnos wiMjiiM «itn duo idc ctpiMiir ani fact i>v \c.ii « -.!« >\
QO.M urwbk' ffc F
FljuiWt l.,hlv(«MrH-frtB">.n# Iff I S '0 R ? b' 50(25,1(121
Ear:y brushing 32" If '0*. 1 ?j 2 K {; \ 51)
FKwrMfc [Jbb'l 39 M 07. 2«l 3-tS1.5,801
Nnncof trwahw* 2^ 1° !0
Raw
Alriciln-Anvilur 3J6 11 (0 8 1 9) 0 9 (0 S, t 1) 40'i»t"i
Early hrwvhmg " 368 2MK\4l| 10(11. 15}
FluMidt' laWc! 130 2 4 (1 :, 4 9) 2 1 (I J, 4 7)
Nan? pi the jtwe 3fi2 10 10
African Ami-wan »* 1. Mi 2, 21) 23{1.». 30)
White* ,»rd i*«» lf«7 10 10
Minid (!98&«questic>n«ibk llwoftfeis) thi-Mju^rc-l •*&?, f™0b gcodnc^vol nt~1^4. lf™.>*t
M9S, P-0000], giwdrwu of fil^THH P«0<4
M.idtl (lW5- i:uon» s) chi-tquate-gj W, "=00001, gwxJnBi of fit"-) 04, !>• 01$
Table 4 summarizes the results from the logistic regression analysis for fluorosis and the
effect of year on race and fluoride exposure. In African- American children who received
fluoride from sources other than water, the risk for very mild to severe fluorosis increased
(OR 1.0 in 1986 vs. 10.5 in 1995), whereas for children of other racial groups there was a
suggestion of slightly decreased risk (OR 0.9). Among those living in fluoridated areas,
the risk for very mild to severe fluorosis increased for both racial groups and was slightly
higher for African- American children (OR 3.9 for African- Americans vs. 2.5 for other
racial groups). The risk for questionable fluorosis did not change from 1986 to 1995 in
non-fluoridated areas for either racial category (OR 1.0); however, there was an increase
in the ORfrom 1986 to 1995 for both racial categories (OR 1.7).
163
January, 2008
-------
American and children oi other r.io*l groups h\ flunj-jde cxpu'tute rai(Xon«*
» ^r^d the •
fluorsd? <
Model I1
Moitel t!J
\ fry
ient for
\ ere
i>
Age group ('1 II ytM
A'cassn American
Year'tab /brush*
018
041
041
i a
044
006
7 14
-on
307
flip::
o *io
0 W
0.000
O.M2
0,049
0,911
0328
0,227
f»M5
I) u40
(irjno
o is
I! 40
i>n
ft*
-0*1
••0.03
-K,6?
0,24
1,26
ttTS
|i W
om
PI58
044,'
0 t2 Hmmei i cinohow gandnrv-oH.l ?utistic-l 2* {/'
* -
-0 Wt)
-sq ure - Y- tt; f*-BMCi <. ilati<-u "0 <>2 Hmmei i cinohow gandnrv-oH.l ?utistic-l 2* {/'
n sn J
* ChiWrosi HI n0RfiuoresS,ih^ ttre3«i
Pt^-rt nl year an Afruan Am^iKan children Umg «n fi«ortdafcd i»vas * OR^«* M** -S»,>KP - ^**t O^o^, ^J^^K ^ ^
fffect ot year on children ^1 other racial groyj,s hvsng m nwondvitn! ^it.'^1, ~ ORv^n M»U v.n< 2*, ORg^^i^^^^- IT
Llfftt tw vtdf an Afrn4n Am«T*4.mi chiMrer uho icLtt\Fed P,uentJi. (ram daily sappktm ht< ut e^riv brushing « b*Hh -
.<*• - v.r»- i!i.ente o!
J and nonflsiomlMt'd ^icasf cu!U!f,^ ^vt'l t'dutation ot
^^n dental Hutmwt*. Kid ^ Rlc.M"^ DMFS of 53
PROFILER'S NOTE: No noteworthy information on the relationship between caries and
fluorosis can be concluded from this data.
STUDY AUTHORS'
CONCLUSIONS:
The two cross-sectional surveys conducted in 1986 and 1995 were considered sufficiently
similar to allow a determination of the changes of risk overtime. The risk of developing
fluorosis did not decline over time in these communities. Water fluoridation affected
fluorosis as evidenced by a significant increase in the prevalence and severity of fluorosis
in Newburgh City. This increase was attributed primarily to the difference in duration of
exposure to fluoridation between the two surveys. Residents of Newburgh City had
continuous exposure since birth in the 1995 survey, but had a 3 year interruption in
exposure in the 1986 survey. However, the increased risk associated with continuous
exposure to water fluoridation may not result in an increase in fluorosis prevalence in
every community after water fluoridation.
In Newburgh Town, fluoridated since 1984, neither the prevalence nor the severity of
fluorosis changed between 1986 and 1995. It is likely that the total fluoride intake did not
change since 71.3% of children in 1986 reported exposure to fluoride through
supplements and/or early tooth brushing.
164
January, 2008
-------
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/date
SJG/3/21/07
PROFILER'S ESTIM.
NOEL/ NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
A higher risk for fluorosis was observed in African-American children, consistent with
other studies. In both surveys, the combined use of daily supplements and early brushing
had the highest odds ratios for very mild to severe fluorosis. There was a significant
association between the use of supplements or early brushing alone with mild to severe
fluorosis.
Although the three-way interaction term (year-race-fluoridation) was not statistically
significant, the between survey risk increase for very mild to severe fluorosis among
children in fluoridated areas was higher for African- Americans. The between survey
change in the effect of fluoride supplements and/or early brushing was much more
dramatic among African-American children. While African American children exposed
to fluoride supplements and/or early brushing were more likely to develop fluorosis in
1995 compared to 1986, the reverse was true for children of other racial groups. No
difference in risk was found for questionable fluorosis among racial groups.
No references or definitions are cited.
Overall, the study was well-conducted and had adequate study design. However, the
study was not designed for development of a dose response to fluoride; the emphasis was
on the prevalence and severity of dental fluorosis in 7 to 14 year old children residing in
fluoridated or non-fluoridated communities. Analysis of surveys conducted in 1986 and
1995 compared the effect of fluoride exposure (via water fluoridation, supplements and/or
early brushing) on fluorosis with respect to year and racial group (i.e., Did the prevalence
and/severity of fluorosis differ over time? Is there a different risk associated with various
sources of fluoride exposure? Are African- American children at higher risk for
developing fluorosis?).
Neither the prevalence nor the severity of fluorosis increased from 1986 to 1995 in
Newburgh Town (after fluoridation) or in any of the non-fluoridated areas. In Newburgh
City (fluoridated), there was a slightly higher prevalence of questionable, very mild, and
mild fluorosis in 1995; these children had lifelong exposure to fluoride in the water while
those in the 1986 survey had a 3 year interruption. For very mild to severe fluorosis,
children using fluoride supplements and early brushing were at the highest risk, although
all fluoride exposures had elevated risk. African- American children were at elevated risk
compared to other racial groups in 1995. No difference in risk was found for
questionable fluorosis among racial groups. Although caries were evaluated (DMFS
score), data was not presented in a clear manner to make any noteworthy conclusions
regarding the relationship between fluorosis or fluoride exposure and caries.
Factors that may affect the results, common to all cross-sectional studies, include:
examiner variation, population differences, representativeness of the sample and recall of
past events for exposure assessment. Fluoride level was not reported for Newburgh
Town. The fluoride concentration in supplement tablets or toothpaste was not reported.
Study design was not suitable for development of a NOAEL for dental fluorosis or caries.
Study design was not suitable for development of a LOAEL for dental fluorosis or caries.
Not suitable (X), Poor ( ), Medium ( ), Strong ( )
While the study was well-conducted, the study design was not conducive to provide data
for a dose-response. The study indicated a higher risk (odds ratio) for mild to severe
fluorosis with fluoride exposure (via water fluoridation, supplements and/or early
brushing). African- American children were at higher risk than other racial groups. The
165
January, 2008
-------
CRITICAL EFFECT(S):
study did not address any issues of plaque or gingivitis.
Prevalence and severity of dental fluorosis
166
January, 2008
-------
Kumar, J.V. and Swango P. A. 1999. Fluoride exposure and dental fluorosis in Newburgh and
Kingston, New York: policy implications. Community Dent. Oral Epidemiol. 27: 171-180.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
Dental fluorosis
Cross-sectional
U.S./New York/Newburgh City (fluoridated water supply): 7-14 yr-old school children
(lifelong residents); 459 (47.9% male; 51.6% African-American) surveyed in 1986; 847
(49.0% male; 41.1% African- American) surveyed in 1995.
U.S./New York/Newburgh Town (fluoridated water supply): 7-14 yr-old school children
(lifelong residents); 289 (58.8% male; 8.3% African-American) surveyed in 1986; 289
(49.1% male, 10.4% African-American) surveyed in 1995.
U.S./New York/New Windsor (non-fluoridated water supply): 7-14 yr-old old school
children (lifelong residents):-; 134 children (52.2% male, 5.2% African-American)
surveyed in 1986; 237 (41.8% male, 6.8% African-American) surveyed in 1995.
U.S./New York/ Kingston (non-fluoridated water supply): 7-14 yr-old school children
(lifelong residents); 425 children (50.3% male; 16.0% African- American) surveyed in
1986; 646 (50.8% male, 19.2% African-American) surveyed in 1995.
U.S./New York/Ulster (non-fluoridated water supply): 7-14 yr-old school children
(lifelong residents), 174 children (50.0% male; 4.6% African-American) surveyed in
1995.
PROFILER'S NOTE: Although the children in New Windsor, Kingston and Ulster did
not have exposure to fluoridated water, some had exposure from fluoride supplements
(tablets) and dentifrices with fluoride (see Table 1 under Exposure Groups).
Children from Newburgh City were exposed throughout their lifetime, except during a
three year period from 1978 to 1981 (affecting children from the 1986 survey as follows:
7 year-olds from 0-2 years of age; 8 yr-olds from 0-3 years of age; 9 yr-olds from 1-4
years of age, etc., ranging up to 14 yr-olds affected from 6-9 years of age).
Children from Newburgh Town surveyed in 1986 were exposed to fluoridation for 2
years. Those surveyed in 1995 were exposed beginning from birth up to age 3 for a
maximum of 1 1 years.
Children from New Windsor, Kingston, and Ulster were not exposed to fluoride in the
drinking water supply.
Table 1 was copied directly from Kumar and Swango (1999) and summarizes the percent
distribution of children in each study community according to age group, gender, race,
and fluoride exposure. Males and females generally were equally represented in all areas,
although there were differences in gender between 1986 and 1995 in Newburgh Town
and New Windsor. In both 1986 and 1995, there were proportionally more African-
American children in Newburgh City compared to other areas. The major (80-95%) race
represented in each community was White or 'other', with 5 to 20% African- American;
the exception was Newburgh City where African- Americans made up 52% and 41% in
1986 and 1995, respectively
167
January, 2008
-------
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN:
Table I, Percentage distribution of children in vudv rommiiratil*. a .tort 11% «. age j?n up RCiKk-r ra«.i *n,I fiuo-idi- f>pw,un;
Vwbu*ghClty Nk'wbmjth lown New WtnJ«>r Kingvlon UKhir*
W,* W5 1-S6 W 1«86 "OS IW, I«S !<»S
Jtomuten stetuv 1= f M- F %!F NF XF MF NT
NumVr «•' s*™ ;(W 2!W !34 2T* 42*' M6 '""*
Age group 7-lf) SI* «* t7» 54- S5S -ill 497 W3 Ml t
'* ** F ;] ,4 4S-- 41! 4 «I 451 ttS IS » 503 40' fc' S
5,,, Mak. 47'> 4"d •«« «t K2 41 S 501 W8 H><>
Ivtpilt 521 SHI 4)2 SO* 47* 58! ->97 492 50 U
RM WhK.'w ,nd othe i 4M 5S « •>! " «",!» «J Mil M« 80S IS 4
Afntun-Aiwrica JH. 417 * 3 1'H 53 h» l<- (I i»2 46
Enposure n«ondjtion4ul d MM 4V J 5K ">
and/or early t ythmg
Hu-iridit.ononl 4">" 508 -4)1
(••«ivtlrsh«nS!U1 "*S - - 41 ; %'." 433 .146 UJ) 348 V-. 1
T,U,,, - 100 76' 3"" MR U* W5 1^5
fv' } fh ' •abtl^ i - - *8 ^ 33 * 32 S V ' 3*> 1 3i S
1 FnfluonrtaWd Ml- npnfl.Kindatcd
2 Children cxammtd m Nc«i»uf)»h City in l'»W> fad an ntiTMp>n r in fli»iri<•!»! tel and/c-r t>m»l-in(; c*t»f on
^ D»I* fcr 198!) mil Biiilibk-
Some children in all the surveyed communities also were exposed to fluoride from other
sources, including fluoride tablet supplements and early brushing (before age 2 years)
with fluoridated toothpaste.
In the non-fluoridated communities 9see Table 1), approximately one third (29-35%) of
the children were not exposed to fluoride from any source considered (water, tablet, early
brushing); approximately another third (3543%) were exposed only by early brushing;
the remainder were exposed via tablet only (4-15%) or tablet plus early brushing (17-
21%).
Data for measuring the fluoride concentrations were not included in the study report.
Water quality parameters were not measured. The fluoride concentration in supplement
tablets or toothpaste was not included.
The purpose of the study was to determine whether the risk imposed by fluoride exposure
has changed during the time period from 1986 to 1995, and to determine the effect of
water fluoridation and other known fluoride sources (fluoride levels not reported) on
dental fluorosis in participating children. Data were obtained from two cross-sectional
surveys, conducted in 1986 and 1995 during the school years in the Newburgh and
Kingston school districts. Analysis was limited to 3500, 7-14 year old lifelong residents
of Newburgh City (fluoridated at 1±0.2 mg/L since 1945, with a 3 year interruption from
1978 to 1981), Newburgh Town (fluoridated since 1984 — level not reported), and New
Windsor, Kingston, or Ulster (all non-fluoridated).
The two surveys were similar in design except for the number of children studied
(n=1307 in 1986; n=2193 in 1995). Dental fluorosis and caries were recorded using
Dean's criteria and the Decayed, Missing and Filled Tooth Surfaces (DMFS) index.
Examination details were not reported, including the examiners, location where
examinations occurred, lighting conditions, or equipment used. Examiners were blind to
residential and fluoridation status of the children. Fluoride exposure data was collected
by questionnaire. The questionnaire design differed between studies as follows:
6) The responses in the 1986 survey were primarily open-ended for the two
questions related to the use of fluoride toothpaste and fluoride tablets. In the
1995 survey, categories for these two items were created.
7) Questions related to school lunch program participation, education of the head of
the household, and breast-feeding were not included in the 1986 survey.
8) The response rate for participation was lower in 1995 (38%) than in 1986 (58%).
168
January, 2008
-------
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
9) Children from the town of Ulster could not be identified from the 1 986 survey
because a question regarding their place of residence was not included.
10) Different examiners were used in each survey, although they were trained by the
same dentist.
PROFILER'S NOTE: Dean's Index of Fluorosis is described in Section 2 of the report).
Dental fluorosis was measured using Dean's classification at the subject level. Fluorosis
was scaled as none, questionable, very mild, mild, moderate or severe. Dental caries was
recorded using the DMFS index (number of decayed, missing and filled permanent tooth
surfaces) and visual-tactile examination.
Regression procedures were used to estimate the effect of fluoridation, fluoride
supplements, and early brushing on dental fluorosis. Analysis involved comparison of
frequency distributions of Dean's fluorosis categories to examine changes over time,
including a "ridit" analysis to examine the changes in severity. The proportion of
children exposed to known fluoride sources was calculated for each category: a)
fluoridation alone; b) fluoridation plus tablet and/or early brushing; c) tablet supplement
alone; d) early tooth brushing alone; e) tablet supplement plus early brushing; f) none of
the above exposures. Adjusted odds ratios and 95% confidence intervals were calculated
for the variables associated with dental fluorosis in the bivariate analysis (p<0. 1). This
model included race and the categories for fluoride exposure variables. The fluoride
exposure variable was introduced as an indicator variable and the reference group
consisted of children from the non-fluoridated areas without known fluoride exposure.
Logistic regression procedures were used to estimate the association between factors and
dental fluorosis separately for the two surveys, and then were used to fit a generalized
"logit" model for examining the effect of year, race and fluoride exposure and their
interactions.
Two separate regression models were constructed to compare categories of fluorosis: one
to compare the questionable category with the normal, and another to compare the very
mild to severe category (combined) with the normal. The examination of two-way and
three-way interaction terms requires larger sample sizes, so exposure categories were
combined as follows: a) fluoridation alone or fluoridation plus tablet and/or early
brushing; b) tablet supplement alone or early tooth brushing alone or tablet supplement
plus early brushing; c) reference group with none of the above exposures. The purpose of
this analysis was to determine whether there were changes in the risk associated with
fluoride exposure between the two surveys for African-American children and children of
other racial groups.
Analyses of crude, covariate-adjusted and ranked DMFS scores were performed to
determine the relationship between caries and dental fluorosis. Four categories of
fluorosis (normal, questionable, very mild, and mild to severe) were created for this
analysis. Other variables included in the model were age, poverty status (based on
participation in free-lunch program) in fluoridated and nonfluoridated areas, presence or
absence of sealants, and education of the head of the household. The dependent variable
(DMFS) was converted to a rank and the tests of significance were based on the rank of
analysis of covariance.
Tables were copied directly from Kumar and Swango (1999). Table 2 shows the
distribution of dental fluorosis according to Dean's index by place of residence and year
of examination. The highest prevalence of the very mild to severe categories was
observed in Newburgh City in 1995 (18.6% combined, vs. 14.8% in Newburgh Town,
14.4% in New Windsor, 11.1% in Kingston, and 14.4% in Ulster). Between-survey
comparisons show that neither the prevalence nor the severity of dental fluorosis
increased after Newburgh Town was fluoridated. The case was the same for non-
fluoridated areas. Changes were evident in Newburgh City (fluoridated), where a ridit
analysis showed that the odds were 4 to 3 that a child examined in 1995 would have at
169
January, 2008
-------
least questionable fluorosis, compared with a similar child in 1986.
Table 2 Dental fluorosfe prevalence in percent .crardtag to Dean's classification by place of residence ad year of exi-mination
Cfean's index
rlKe %„ Status Number Norm.) Questionable Very mild Mild Moderate Severe^
„, t t.r--> iaa F 4"» 784 13.7 4-8 2,2 0.9
Newburgh City iwo r ***>? iu^*
J9J51 F 847 «•* 18-3 *2-8 w *Jl<*
Ncwburgh Town 1986 NF 2® 73.0 13,1 8,7 4.2 1-0
New Windsor V» NF 134 76-1 9-7 « « «
1995 NF 237 7S.5 10,1 8-9 S.I 0.4
Kini«« 1*4 NF 425 81.4 11.3 4.3 2.1 0,7
IW5 NF M6 | 81.4 M '-' J-' U'J
Ulster 1986 NF - -^ 98 2~9 fl 06
' Ridit tot this group «« 0.58 relative to 0.50 tor the 1W survey (statistically apiificanl, F<0.05). AH ofcr betwcen-survey
comparisons vieldtd ridit values o< less than 051 . .
' The pimalence of very mild to severe dental fluorosis among children bom after the ,mplemerrtat.on of fluoridation m
Ncwburgh Town was 14.7 (37/252).
Table 3 summarizes the adjusted odds ratios associated with fluoride exposure and race
by year of study. Children using fluoride tablets and early brushing had the highest odds
ratios for very mild to severe fluorosis in both 1986 (OR: 5.0; CI: 2.5, 10.2) and 1995
(OR: 4.0; CI: 2.4, 6.9). Elevated odds ratios were observed for all the fluoride exposure
variables in both years; however, exposure to fluoridation alone in 1986 was not
statistically significant. African-American children studied in 1995 were at higher risk
(OR: 2.3; CI: 1.8, 3.0) for dental fluorosis than children of other racial groups. While
elevated odds ratios were observed for questionable fluorosis in 1995 (range: 1.6 to 4.4),
they were not statistically significant in 1986 (range: 1.0 to 1.5).
JMf J (Md-i tatnos wiMjliM mtn fluoride ctfiMiir ani race cv \c.it iX '.Hi '\
»>\t i rwWe flu i-i-ws W<\ rmlj to MX ««•
- "
FluoiirJh.rtm-cirh t-rush'pt> or 1 \Wi-l 231 11 (OR 231 2 I (1 0, 4 4>
1 luor.iloli,™ lUiiw KA 1 5 (0 6, 1 re*ing or tablet 586 41 ;: 6. 72} 3 3 (1.152)
FluoiHlatMt S!ai58l 25053")
nuoridv labk-i* early brushing 19? ."'xO'i^Si 40'i»t"i
Early hrwvhmg 3« 2. 5 (13. 411 10(11. 15}
FluMidt' laWc! 130 > 4 (1 2. 4 9} 2 <1 (I J, 4 ~)
None pi the jtwe 3K2 1.0 10
Atriun Araitican f > 1 M 1 2. 2 n 2 3 (1 .». 3 0)
Vth^e* ,*rtj t*tos^ i^7 HI 10
Miniei (!98&^]uestic>fl«ible (lltOftfeis) thi™M^uate~4 36*4, P™0b g! lit^l^C n~ ^%
M«Hii'I ((Wft-vt-ry mild-sevetv DUOITOS) chl square -2(>95 P-0 0001, giiodiwu of ht»7 KK, T=0 14
Mxkl ClWS-vi-ry raiM-«ve«> i:uon» s) chi-squate'lUM l'=00001, K«xlt>is5. of fit- 1 04, P-098
Table 4 summarizes the results from the logistic regression analysis for fluorosis and the
effect of year on race and fluoride exposure. In African- American children who received
fluoride from sources other than water, the risk for very mild to severe fluorosis increased
(OR 1.0 in 1986 vs. 10.5 in 1995), whereas for children of other racial groups there was a
suggestion of slightly decreased risk (OR 0.9). Among those living in fluoridated areas,
the risk for very mild to severe fluorosis increased for both racial groups and was slightly
higher for African- American children (OR 3 .9 for African- Americans vs. 2.5 for other
racial groups). The risk for questionable fluorosis did not change from 1986 to 1995 in
non-fluoridated areas for either racial category (OR 1.0); however, there was an increase
in the ORfrom 1986 to 1995 for both racial categories (OR 1.7).
170
January, 2008
-------
American and children oi other r.io*l groups h\ flunj-jde cxpu'tute rai(Xon«*
» ^r^d the •
fluorsd? <
Model I1
Moitel t!J
\ fry
ient for
\ ere
i>
Age group ('1 II ytM
A'cassn American
Year'tab /brush*
018
041
041
i a
044
006
7 14
-on
307
flip::
o *io
0 W
0.000
O.M2
0,049
0,911
0328
0,227
f»M5
I) u40
(irjno
o is
I! 40
i>n
ft*
-0*1
••0.03
-K,6?
0,24
1,26
ttTS
|i W
om
PI58
044,'
0 t2 Hmmei i cinohow gandnrv-oH.l ?utistic-l 2* {/'
* -
-0 Wt)
-sq ure - Y- tt; f*-BMCi <. ilati<-u "0 <>2 Hmmei i cinohow gandnrv-oH.l ?utistic-l 2* {/'
n sn J
* ChiWrosi HI n0RfiuoresS,ih^ ttre3«i
Pt^-rt nl year an Afruan Am^iKan children Umg «n fi«ortdafcd i»vas * OR^«* M** -S»,>KP - ^**t O^o^, ^J^^K ^ ^
fffect ot year on children ^1 other racial groyj,s hvsng m nwondvitn! ^it.'^1, ~ ORv^n M»U v.n< 2*, ORg^^i^^^^- IT
Llfftt tw vtdf an Afrn4n Am«T*4.mi chiMrer uho icLtt\Fed P,uentJi. (ram daily sappktm ht< ut e^riv brushing « b*Hh -
.<*• - v.r»- i!i.ente o!
J and nonflsiomlMt'd ^icasf cu!U!f,^ ^vt'l t'dutation ot
^^n dental Hutmwt*. Kid ^ Rlc.M"^ DMFS of 53
PROFILER'S NOTE: No noteworthy information on the relationship between caries and
fluorosis can be concluded from this data.
STUDY AUTHORS'
CONCLUSIONS:
The two cross-sectional surveys conducted in 1986 and 1995 were considered sufficiently
similar to allow a determination of the changes of risk overtime. The risk of developing
fluorosis did not decline over time in these communities. Water fluoridation affected
fluorosis as evidenced by a significant increase in the prevalence and severity of fluorosis
in Newburgh City. This increase was attributed primarily to the difference in duration of
exposure to fluoridation between the two surveys. Residents of Newburgh City had
continuous exposure since birth in the 1995 survey, but had a 3 year interruption in
exposure in the 1986 survey. However, the increased risk associated with continuous
exposure to water fluoridation may not result in an increase in fluorosis prevalence in
every community after water fluoridation.
In Newburgh Town, fluoridated since 1984, neither the prevalence nor the severity of
fluorosis changed between 1986 and 1995. It is likely that the total fluoride intake did not
change since 71.3% of children in 1986 reported exposure to fluoride through
supplements and/or early tooth brushing.
171
January, 2008
-------
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/date
SJG/3/21/07
PROFILER'S ESTIM.
NOEL/ NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
A higher risk for fluorosis was observed in African-American children, consistent with
other studies. In both surveys, the combined use of daily supplements and early brushing
had the highest odds ratios for very mild to severe fluorosis. There was a significant
association between the use of supplements or early brushing alone with mild to severe
fluorosis.
Although the three-way interaction term (year-race-fluoridation) was not statistically
significant, the between survey risk increase for very mild to severe fluorosis among
children in fluoridated areas was higher for African- Americans. The between survey
change in the effect of fluoride supplements and/or early brushing was much more
dramatic among African-American children. While African American children exposed
to fluoride supplements and/or early brushing were more likely to develop fluorosis in
1995 compared to 1986, the reverse was true for children of other racial groups. No
difference in risk was found for questionable fluorosis among racial groups.
No references or definitions are cited.
Overall, the study was well-conducted and had adequate study design. However, the
study was not designed for development of a dose response to fluoride; the emphasis was
on the prevalence and severity of dental fluorosis in 7 to 14 year old children residing in
fluoridated or non-fluoridated communities. Analysis of surveys conducted in 1986 and
1995 compared the effect of fluoride exposure (via water fluoridation, supplements and/or
early brushing) on fluorosis with respect to year and racial group (i.e., Did the prevalence
and/severity of fluorosis differ over time? Is there a different risk associated with various
sources of fluoride exposure? Are African- American children at higher risk for
developing fluorosis?).
Neither the prevalence nor the severity of fluorosis increased from 1986 to 1995 in
Newburgh Town (after fluoridation) or in any of the non-fluoridated areas. In Newburgh
City (fluoridated), there was a slightly higher prevalence of questionable, very mild, and
mild fluorosis in 1995; these children had lifelong exposure to fluoride in the water while
those in the 1986 survey had a 3 year interruption. For very mild to severe fluorosis,
children using fluoride supplements and early brushing were at the highest risk, although
all fluoride exposures had elevated risk. African- American children were at elevated risk
compared to other racial groups in 1995. No difference in risk was found for
questionable fluorosis among racial groups. Although caries were evaluated (DMFS
score), data was not presented in a clear manner to make any noteworthy conclusions
regarding the relationship between fluorosis or fluoride exposure and caries.
Factors that may affect the results, common to all cross-sectional studies, include:
examiner variation, population differences, representativeness of the sample and recall of
past events for exposure assessment. Fluoride level was not reported for Newburgh
Town. The fluoride concentration in supplement tablets or toothpaste was not reported.
Study design was not suitable for development of a NOAEL for dental fluorosis or caries.
Study design was not suitable for development of a LOAEL for dental fluorosis or caries.
Not suitable (X), Poor ( ), Medium ( ), Strong ( )
While the study was well-conducted, the study design was not conducive to provide data
for a dose-response. The study indicated a higher risk (odds ratio) for mild to severe
fluorosis with fluoride exposure (via water fluoridation, supplements and/or early
brushing). African- American children were at higher risk than other racial groups. The
172
January, 2008
-------
CRITICAL EFFECT(S):
study did not address any issues of plaque
or gingivitis.
Prevalence and severity of dental fluorosis
173
January, 2008
-------
Leake, J., F. Goettler, B. Stahl-Quinlan and H. Stewart. 2002. Has the level of dental fluorosis
among Toronto children changed? J Can Dent Assoc 68(l):21-5.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
Dental caries and fluorosis (maxillary permanent anterior teeth)
Prevalence study of dental caries and fluorosis.
Canada/Toronto: The current study reports results for 2435 children aged 7 to 13 years out
of an overall total of 3657 children (number of males and females not specified) aged 5, 7 or
13 years from high-, medium- and low-risk schools in each of four health regions (3x3x4
cells) of Toronto, Canada. (According to Toronto standards, a school is at medium risk if
between 9.5% and 14.0% of junior and senior kindergarten children have decay on 2 or
more teeth. Definitions for low- and high-risk schools were not provided in the study
report.)
None.
Study was conducted in 1999-2000 with 7 and 13 yr-old children (N = 2435 in these age
classes).
The study population was not divided into groups based on fluoride levels in the drinking
water. Fluoride concentrations in drinking water were not measured, although the authors
report that the Toronto water system has been fluoridated since 1963, and the F
concentration was gradually reduced from 1.2ppmto0.8ppmin 1999 to meet revised
Canadian water standards (Health Canada 1996).
No estimates of fluoride intake were provided.
None
The purpose of the study was to obtain valid estimates of the oral health status of a
probability sample of children in 4 regions of the newly amalgamated city of Toronto,
Ontario, Canada. An overall total of 3657 children (number of males and females not
provided) were enrolled in the parent study, of which 2435 between the ages of 7-13 were
evaluated in the current report. The current study was conducted during the 1999-2000
school year and included examinations for caries in all children and for fluorosis of the
maxillary permanent anterior teeth in 7- and 13-year-old children. One of two specially
trained dental hygienists examined each child's teeth and periodontal tissues. The two
examiners were trained during separate, day-long sessions by the senior investigator
(J.L.L.). Compliance with the criteria of the protocol was rechecked approximately
biweekly by the senior investigator, who independently examined children enrolled in the
study and compared his results to those of the 2 examiners.] A two-stage sampling process
was used. During the first stage, 6 schools for each cell (age, risk category and region) were
selected. Then a random start and cell-specific sampling ratio (age-specific enrolment in the
6 schools divided by 100) was used to select the children to be included in each cell.
Informed consent was obtained from parents. The survey examination followed the protocol
issued by the Ontario Ministry of Health (Ontario Ministry of Health, 1998), as described
under PARAMETERS MONITORED. The protocol also calls for examiners to indicate
whether the child has urgent treatment needs; criteria include the presence of pain, infection,
hemorrhage, trauma, large open lesions and acute periodontal conditions.
The survey examination followed the protocol issued by the Ontario Ministry of Health.
The protocol states that only dentinal caries are to be scored at the level of the tooth, i.e.,
surface scores are not recorded. Fluorosis was measured on the maxillary permanent
anterior teeth of 243 5 children aged 7 and 1 3 years according to the Tooth Surface Index of
174
January, 2008
-------
Fluorosis (TSIF; see Section 2). The ministry protocol states that TSIF be scored, in terms
of the highest score on bilateral pairs of teeth, as none (TSIF=0); parchment white patches
visible on less than one-third of the tooth surface (TSIF=1); parchment white color visible
on at least one-third but less than two-thirds of the tooth surface (TSIF=2); parchment white
color visible on two-thirds or more of the tooth surface (TSIF=3); and staining or pitting (or
both) in conjunction with a TSIF score of 1, 2 or 3 (TSIF=4).
STATISTICAL METHODS:
Data were transferred to the Statistical Package for Social Sciences (SPSS). The descriptive
findings were weighted according to the city's population in each age group. Tests for
associations with potential determinants were conducted on the unweighted data. Basic
findings were recorded according to the O'Keefe template (O'Keefe, 1995). For fluorosis,
the reporting cut-off of a score of 2 or more reflects the untested hypothesis that most
parents and children would not be aware of a condition scoring 1. No fluorosis findings
were reported in 5-year-olds since only permanent teeth were examined for this condition.
RESULTS:
Results of the study in Tables 1 through 4 are shown directly from Leake, et al. (2002) and
are limited to findings of fluorosis on maxillary permanent anterior teeth and its relationship
to prevalence/severity of caries. The authors state that "a score of 1... is by convention
assumed to be aesthetically unimportant" (p. 23).
Table 1 Caries and flunrosis in Toronto
children (weighted findings of the
2IHH) Denial Indices System survey)
Age- group'
weighted % of
Indicator
/-year-olds 13-yejr-olds
(weighted weighted
n = 2792,1 ?i = 2493.1
h curies expeneno?
iJigent treatment needs
With 2 or more decayed teeth
Mean deft + DMFT land SD\
Mc-an DMFT (and SO)
With moderate fluorosis (TSIF ^ 2)
41.3
7,4
7,0
1,59(2,1)
14.0
39.3
1.7
2,0
.13(2,0)
12,3
a£xo?pt
Sf1 = i'
indicate
'uto/i, 7S'C
Surface Index of Fiuorosis,
175
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
Table 2 Distribution of TSIF scores (weighted
findings of the 2000 Dental Indices
System survey)
Age giotip
weighted ">. of iubject
TSIF score 7-veai-old:, 1 3-yeai-olcK
i weighted weighted
n = 2702) ,n = 24931
0 fno fluorosni 73,2 70 &
1 tflucro'ji; on less than
one-third of the touthi 1 2,8 8,2
2 ifluorosii on af loast
one-thud but le;s trun
tvvo-thirdo of the toothi 9.2 6,6
3 i flue rose, en tao-thiid1,
ur moit of fhe toothi 4,5 3,9
4 (Staining pitting or hath,
in conjunction with
TSIF scene of 1 2 or ?,) 0,3 1.8
TSi'F= loom Surface t'/idet of fluarosis.
Table 3 TSIF scores hy birthpl«nt» iinuint* 7- and IM-year-old
participants (weighted findings uf tlic- 2IHIO Denial Indices
System «mrey)
Birthplace;
% of subjects
TSIF score Toronto, Elsewhere Outside Not Total p value"
Ontario in Canada Canada stated (n = 2435)
(n = 1285) (n = 61) (n = 800) (n = 309)
0 73,4 85,2 86,9 63,8 76,9
>1 26,6 14.8 13,1 38,2 23,1 . 0,001
>2 15,4 8,2 4,9 23,0 12.7 * 0.001
•CW-square test
TSIF = Tooth Surface Index of Fluorosls.
Table 4 Kfkitiomhip botvwen woritv of fluorosis and tarios
experience amoii|» 7-jvar-oIds (weightpd undines of the
2000 Dental Indices System snney)
Carfet expeiienc* TSIF = 0 TSIF = 1 TSIF > 2 All scoics p value
|f) = 9«) In = 146) In = 162) {n = 1210)
"o of children s 424 30.1 37.0 40,2 nni4»
^Apenente tdpft + DMFT ^ !<
MP tn iSeft + DMFT 1,69 1.36 1,23 1.59 O.Q&T*
""Cfii-^qu ire te-t 2 rt
-------
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/Date
VAD/01-01-
07
PROFILER'S ESTEM.
NOAEL
PROFILER'S ESTEM.
LOAEL
SUITABILITY FOR DOSE
RESPONSE MODELING
CRITICAL EFFECTS:
Interpretation of results was limited by the Dental Indices System protocol, which provided
information only on place of birth and age to be used as possible variables for levels of
fluorosis.
Examiner biases could have resulted because examiners may have under-reported caries in
children they previously identified as having fluorosis since fluorosis incidence was
recorded before caries incidence in the protocol.
Ontario Ministry of Health. 1998. Dental index system. Toronto, Canada: Public Health
Branch.
O'Keefe, J.P. 1995. A template dental health status report for Ontario public health units.
Can J Community Dent 10(1)18-24.
Health Canada. 1996. Guidelines for Canadian Drinking Water Quality — Supporting
Documents. Ottawa, Canada, (http://www.hc-
sc.gc.ca/ehp/catalogue/bch_pubs/dwgsup doc/fluoride.pdf)
The reported study was a good measure of the prevalence of caries and fluorosis in Toronto
children aged 7 and 13 during the 1999-2000 school term. Birthplace information was not
available for all children and if available, may not have provided accurate exposure
information during the susceptible period for fluorosis development on the permanent teeth.
One hygienist examined 90% of the children, which could have biased the results.
Although the results indicated that a small percentage of the study group exhibited severe
fluorosis, the effect of severe fluorosis on caries prevalence was not evaluated, other than to
document that of 37% of the 7 yr olds with "moderate" fluorosis (TSIF > 2) had caries vs.
30. 1% for those with only "mild" fluorosis (TSIF = 1; Table 4). Only 0.3% of the 7 yr-olds
and 1.8% of the 13 yr-olds had TSIF scores of 4 (Table 2).
REVIEWER'S REMARKS: Although not reported in the current study, it is presumed that
the drinking water to which study participants were exposed was in compliance with Health
Canada (1996) requirements (e.g., 1.2 ppm prior to 1999 [See the previous page.] and
reduced to 0.8 ppm in 1999).
The study design did not estimate fluoride intake and the association with fluorosis.
The study design did not estimate fluoride intake and the association with fluorosis.
Not suitable (X_); Poor (_); Medium (); Strong (_)
Dental caries and fluorosis (maxillary permanent anterior teeth)
177
January, 2008
-------
Levy, S.M., Warren, J.J., Broffitt, B., Hillis, S.L. and Kanellis, M.J. 2003. Fluoride, beverages and
dental caries in the primary dentition. Caries Res. 37:157-165.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
Dental caries in children up to 6 years
Longitudinal cohort of children whose parents were recruited at the time of the child's birth
during 1992-1995.
PROFILER'S NOTE: There are 4 articles that report on different findings but all use the
population of children that originally were part of the Iowa Fluoride Study (IPS) that
occurred during 1992 to 1995 and was a birth cohort study of fluoride exposures and intake,
fluorosis and caries. The profiler did not put all assessments on the same profile as the data
were sufficiently different to combine. The reports were as follows: Levy et al, 2003;
Marshall et al., 2004; Levy et al., 2006a and Levy et al., 2006b.
US/Iowa: 29 1 individuals were recruited from eight Iowa hospitals to participate in the Iowa
Fluoride Study.
No control population was included.
Children were followed until between the ages of 4 and 6 years.
The participants were not divided into exposure groups. Although weighted average fluoride
levels in water sources were placed in 3 groups: < 0.3 ppm (10.9%); 0.3-0.6 ppm (16%) and
X).6 ppm (73. 1%); groups were not compared for fluorosis prevalence. Using questionnaires
at scheduled intervals, dietary and non-dietary sources of fluoride were assessed to estimate
intake.
Daily fluoride exposure, fluoride intake and categories of dietary intake were determined
at the age of the child for each questionnaire response. Linear interpolation was used to
obtain daily estimates of values for intervals between returned questionnaires. Finally,
yearly summaries of fluoride and dietary intake for each subject were obtained by
averaging the daily values (observed and interpolated) over each of the first 4 years of life.
The questionnaires asked parents to summarize their child's dietary intake for the previous
week. Dietary intake was partitioned into eleven broad categories: water, formula, breast
milk, cow's milk, juices and juice drinks, non-juice beverages as purchased, beverages
made from frozen concentrates, beverages made from powdered concentrates, ready-to-
feed baby food, infant cereal made from powder, and other foods made with water (Jell-
O®, soup, etc.). Formula and juice drink questions contained additional detail to ascertain
the amount of personal water that was added to each beverage since subjects sometimes
used ready-to-feed juice or formula but at other times used products made from powder or
liquid concentrate. Fluoride from dentifrice was estimated by combining frequency of use,
brand-specific fluoride concentration and estimates of the amounts of dentifrice used and
ingested at each brushing for each time period beginning at 6 months of age. Estimated
daily fluoride intake from dietary fluoride supplements similarly was determined by
combining frequency of use with brand-specific dosage information separately for each
time period. Estimates of fluoride intake from water incorporated both the daily amount of
water consumed by the child and an estimate of the fluoride concentration (parts per
million) in each of the major water sources used by the child (e.g. home, child care).
The method for analyzing fluoride levels was not described. Water fluoride levels were
determined through assay of individual wells or filtered public water supplies, assays of
commercial bottled waters, and documentation of fluoride levels for public water supplies.
Parents were recruited at the time of their child's birth from eight Iowa hospitals.
Demographic information was obtained at the time of recruitment. Using questionnaires
sent at scheduled intervals, parents also provided information about the children's water
178
January, 2008
-------
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental caries
sources, beverage and selected food intake, use of dietary fluoride supplements, and
toothbrushing habits. Questionnaires were sent out at regular intervals starting at 6 weeks
of age, with the greatest frequency during the child's first year of life (5 times) and with
decreasing frequency thereafter. Dental examinations were conducted on primary
dentition when the children were between 4 and 6 years.
Oral examinations of children were conducted to determine dental caries in the primary
dentition according to the criteria of Warren et al. (2002). Children were examined at age
4-6 years, using did2_3f criteria that differentiated between non-cavitated (d^ and
cavitated (d2-s) carious lesions with each surface scored as sound, filled, or as a cavitated
or non-cavitated lesion.
The Statistical Analysis System (SAS) Version 8 was used for data analysis. All predictor
variables (except gender) were transformed to 3 -level ordinal variables having the
following values: 0 = low, 1 = medium, and 2 = high. Assignment to the low, medium,
and high levels was based on distribution percentiles, so that roughly equal numbers of
subjects fell into each level for each transformed variable. The 3-level variables were then
used as linear predictors in logistic regression models. Odds ratio estimates in the logistic
regressions should be interpreted as the estimated change in the odds ratio resulting from
an increase to the next higher category (low to medium, or medium to high) in the
corresponding variable. The association between certain variables and caries experience
was determined by computer model. Subsets of variables included in the model were
none, age and gender, water consumption, and sugar beverages/milk consumption.
Only 23% of the children had caries experience and approximately 73% had weighted
average fluoride levels in water sources that exceeded 0.6 ppm (Table 1).
A logistic regression model was used to calculate odd ratios in order to show relationships
between the parameter variables used in the model. Water consumption (36-48 months),
more frequent toothbrushing (36-48 months), and milk consumption (24-36 months) were
statistically significantly associated with decreased odds ratios for caries. In contrast,
consumption of sugar beverages or milk (6 weeks to 12 months) was statistically
significantly associated with an increased odds ratio for caries.
179
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date
REMARKS CSW
1/9/2007
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM.
LOEL/LOAEL
POTENTIAL
SUITABILITY FOR DOSE-
RESPONSE MODELING:
CRITICAL EFFECT(S):
Table 1 . Characteristics of the sample (n = 29 1 )
Variable Category Percentage
Child's sex male 46.7
female 53.3
Mother's age 17-24 8.9
25-29 37.8
3O-34 31.6
35-45 2 1 .7
Mother's education up to high school 14.8
some college 34.0
college graduate or more 51.2
Family income
-------
Levy, S.M., J.J. Warren, B. Broffitt, and M.J. Kanellis. 2006a. Associations between dental
fluorosis of the permanent and primary dentitions. Journal of Public Health Dentistry. Vol.
66, No. 3, p. 180-185.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
Dental fluorosis in primary and permanent dentition
Birth cohort; longitudinal study.
PROFILER'S NOTE: There are 4 articles (Levy et al, 2003; Marshall et al, 2004; Levy et
al, 2006a and Levy et al., 2006b) that report on different findings but all use the same
population of children that were originally part of the Iowa Fluoride Study (IPS) during
1992 to 1995. The IFS was a birth cohort study of fluoride exposures and intake, fluorosis
and caries. The profiler did not place all assessments on the same profile, as the
individually reported data were sufficiently different.
US/Iowa: 601 children who participated in the Iowa Fluoride Study; their parents were
recruited in the hospital at each child's birth during 1992-1995. Institutional Review Board
approval, parental consent and child assent were all obtained.
PROFILER'S NOTE: The applicability of this study population for the general United
States is somewhat limited as the subjects and their families were mostly Caucasian,
middle-income and the majority of the parents had a college education.
None described
Parents of the children were recruited at birth from 8 Iowa hospitals during 1992-1995.
The children were given dental examinations to identify fluorosis at ages 4-6 (mean age
5.2) and 7-12 (mean age 9.2).
The IFS database was used to determine all possible routes of fluoride exposure.
Information in the database included drinking water fluoride levels; filtration sources;
water, beverage and selected food intake; use of fluoride supplements and fluoride
dentifrice; and body weight.
PROFILER'S NOTE: Although this study stated that fluoride from all the above
mentioned sources were collected, no quantitative values were provided.
All possible sources of fluoride exposure, including food and drink, were evaluated by
questionnaire.
Analytical methods for determining fluoride concentrations were not described in the
report.
The population was a birth cohort of children participating in the Iowa Fluoride Study.
Children were given two dental examinations for fluorosis. One occurred between 4-6 years
old and the second between 7-12 years old. Examinations were performed by two trained,
calibrated dentists with portable equipment and halogen headlights. Additional
examinations were conducted by both examiners on a subset to assess inter-examiner
reliability.
Parents were sent questionnaires five times during the first year and then 2-3 times per year,
thereafter, to assess fluoride intake. Questionnaires addressed water sources, filtration
status, water, beverage, and selected food intake, use of dietary fluoride supplements and
dentifrice and body weight. Total dietary fluoride intake was divided by body weight for
each questionnaire. Average daily fluoride intake (mgF/kg bw) was estimated from birth to
36 months and again from 36 to 72 months using the trapezoidal method of calculation for
area-under-the-curve (AUC).
181
January, 2008
-------
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
PROFILER'S NOTE: Although the paper describes the sources of fluoride, no quantitative
data on the actual concentration of fluoride from each source were included.
At the first examination at ages 4-6, primary teeth were assessed for fluorosis using the
Tooth Surface Index of Fluorosis (TSIF), adapted for primary teeth (Warren et al., 2001).
For the second examination, the Fluorosis Risk Index (FRI) (Pendrys, 1990) was used to
evaluate the early erupting permanent teeth (8 permanent incisors and 4 first molars) and
the TSIF was used for primary second molars. The FRI was selected by the study authors,
as it scores fluorosis on four zones of the tooth. For this study, three FRI zones of each
buccal surface (incisal edge/cusp tip, incisal/occlusal third and middle third, with gingival
zones excluded due to less full eruption) were included. The permanent teeth were
categorized as follows: 1) definitive cases: at least one FRI score of 2 (white striations) or 3
(staining/pitting/deformity) on more than one-half of a surface zone; 2) questionable: with a
maximum FRI score of 1 for less than half of a zone clearly or possibly affected by white
striations; and 3) none: all zones scored as FRI=0 (no indications of fluorosis) or 7 (non-
fluoride opacity).
PROFILER'S NOTE: The NRC (2006) states that the FRI index by Pendrys (1990) was
specifically designed for use in case-control studies, of which there are very few.
Statistical analysis was performed using SAS version 9 (SAS 2003). Both permanent
incisor and permanent first molar fluorosis results were separately assessed at age 5 and at
age 9 using relative risks and logistic regression. Relative risks and 95% confidence
intervals were calculated according to the SAS cohort study method. Two separate logistic
regressions predicting definitive permanent incisor fluorosis used 0-36 and 36-72 month
AUC fluoride intake, respectively, in addition to primary second molar fluorosis at age 5.
Two additional logistic regressions predicting definitive permanent first molar fluorosis
also used primary second molar fluorosis assessed at age 5, as well as 0-36 and 36-72
month AUC fluoride intake, respectively.
Tables 1 and 2 are copied directly from Levy et al. (2006a). The tables provide the
relationships between permanent incisors and molars as compared to primary molar
fluorosis. Table 1 shows a significant relationship between age 5 and age 9 primary tooth
and permanent incisor fluorosis, although they were stronger for age 9 primary second
molars. The age 5 primary tooth fluorosis prevalence rates were 2.2% for the first molars
and 9.8% for the second molars. Prevalence rates for fluorosis of the permanent incisors
(age 9) were 36.3% definitive, 27.3% questionable and 36.4% none, while fluorosis
prevalence for the permanent first molars was 20.0% definitive, 25.5% questionable and
54.6% none. Almost all dental fluorosis was mild, with only 8 individuals (1%) with
moderate (dark staining)/severe (pitting) permanent tooth fluorosis (FRI score of 3) and
only 2 (0.3%) with severe primary tooth fluorosis (TSIF score of 5).
182
January, 2008
-------
Fluoride intake and
prediction of fluorosis
Table 1
Relationships between permanent incisor and primary molar flaorosis
Percentage with Relative Risk for
Primary Mote Primary Tooth n Permanent Incisor Ryorosis Definitive Fluorosu
Fluorosis Huorosis (vs. Questionable/ None)
Definitive Questionable None HR 95% CI
AgeS Yes 2 13 85 8 8 2.4 1.9-3.1
1" Molar No 98 588 ,15 28 37
AgeS Yes 10 59 76 12 12 2.4 2,0-2.9
2rf Molar* No 90 542 32 2.9 39
AgeS Yes 13 80 8! 8 11 2.8 2.3-3,3
2°' Molar No 87 52) 2* 10 411
*AH 13 subjects wjSh primarv firs? rcwtar Huonw* »dsu had primary swund moUr fluw^
Table 2
Relationships between permanent first molar and primary molar fluorosis
Pt-rctnUgi with Relative Risk for
Prunarv Molar Primary Tooth n Permanent First Molar FIutircKis {",',) Ochniatt: Huorusis
Fluiirosis Huitrusis (vs, Questionable/None)
Definitive Qiu>&boiubl« None RR **51 Ci
AgeS Its 2 H 77 V, 0 4.1 2 12 JJ< 2S-50
2«* Molar* Ni> «»0 542 If, Z=i 54
Age" Yes 13 80 61 24 }* 45 3.4-5 "
2" Molar No 87 S21 14 26 M
*AH !$ siubsct.ls with pnmarv hrtt molar fluorosis also Kad pnman. mtwrtJ misij(iT fluoro*-^
PROFILER' S NOTE: While the tables do support that fluorosis in primary teeth can be a
good predictor that fluorosis will be observed in permanent incisors, the degree of fluorosis
observed is not provided. There is no reported evidence to support any dose-related trend
(i.e. as fluorosis in primary teeth was more severe, the fluorosis in permanent teeth was
also). The study states only that a majority of the children had mild fluorosis, which is not
considered adverse. Although the study collected data on the sources of fluoride (i.e. food,
water, beverages, supplements), data are not included to identify which sources are most
important, or to establish relative source contribution.
Figures 1 and 2 are copied directly from Levy et al. (2006a) and show logistic regression
predictions of permanent incisor and molar fluorosis based on primary 2nd molar incidence
of fluorosis and the intake of fluoride from 0-36 months. The graphs show that those
children with primary molar fluorosis were much more likely to have permanent incisor
fluorosis (76% vs. 32%) and permanent molar fluorosis (59% vs. 16%) at all levels of
fluoride intake. Similar results were found when the 36-72 months results were plotted, but
the data were not provided in the reported study.
183
January, 2008
-------
Logisie Rsgntssien Prediction of Permanent Fluorosfs
Primaiy 2<*> Malar f lyoroste indicator
and AUC Fluoride from Ag» 0 to 38
N=343
1,i
OJ9-
Pfljbeflilily
of as-1
Permanent rj,4 -
incisor 03j
0.2 <
w
0$ J
MmqrlMll
•NaMmylMir
Fluarosls
Ofl§ ftffi 0,04 OJI 0,(B tt10 0,12
AUC Ftyondo (Ateks ai Afl« 0-38 Montis
few]
QiJiBSfiJtSkCli
12 (1114)
Fbcrjde Intake
Figure 2
LogiSlB Regrwsiion Prediction of Permanent First Molar Flworosis
Using Primary 2M Molar Fluowas Indicator
and AUC Flyonde intake from Age 0 to 36 months
0.7
EsJinwSaiJ na
PwbBWIIly
Of 05
Pfifmjnerrt
First Molar
Fluomsis
ntamiyMohr
No Primary MoJar
Ruemtl*
0.00 0.02 0.04 0.06 0.08 0.10 0,12
AOC Fluonde tntaks at Ag« 0-3fi Months
B«jrltl»tm*»|001 Ti
7,S (3-8,16,8)
1.J (12,1.5)
PROFILER'S NOTE: Although fluoride intake levels were provided, the parameters for
the degree of fluorosis were not included on the chart; the study only states that the
majority of the children had mild fluorosis, which is not considered adverse. Although the
study collected data on the sources of fluoride (i.e. food, water, beverages, supplements),
data are not included to identify which sources are most important.
STUDY AUTHORS'
CONCLUSIONS:
There is a strong association between primary and permanent tooth fluorosis that is
independent of the level of fluoride intake. The detection of primary tooth fluorosis in pre-
school children should alert clinicians and parents to the high likelihood of subsequent
fluorosis in the permanent dentition.
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
Warren, J.J., S.M. Levy andMJ. Kanellis. 2001. Prevalence of dental fluorosis in the
primary dentition. Journal Public Health Dent. Vol. 61(2), p. 87-91.
PROFILER'S
Initials/date
While the Iowa Fluoride Study appears to be a thorough study, this paper does not include
184
January, 2008
-------
REMARKS
DFG/1-07
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM.
LOEL/LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
data critical for determining a dose-response. Levy et al. (2006a) provided limited data on
the fluoride concentration intake levels to which the children were exposed, but did not
provide any quantitative data on the sources of fluoride or relative source contribution.
Limited data on the degree of fluorosis observed indicate that very few children
experienced severe (adverse) fluorosis. The authors also use a Fluorosis Risk Index
(Pendrys, 1990) that is of limited applicability according to the NRC (2006). The study
population was not a true reflection of the majority of children in the United States as most
were from families within the middle-income bracket with parents having a 4-year college
degree which tends to indicate better preventative dental care was available.
Data are unsuitable for determining a dose-response for fluorosis.
Data are unsuitable for determining a dose-response for fluorosis.
Not suitable (X), Poor (_), Medium (J, Strong (J
Dental fluorosis in permanent dentition as related to its presence in primary dentition.
185
January, 2008
-------
Levy, S.M., L. Hong, J. J. Warren and B. Broffitt. 2006b. Use of the fluorosis risk index in a cohort
study: the Iowa fluoride study. Journal of Public Health Dentistry. Vol. 66, No. 2, p. 92-96.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL
POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL
METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
Dental fluorosis
Cohort
PROFILER'S NOTE: There are 4 articles that report on different findings but all use the
population of children that originally were part of the Iowa Fluoride Study (IFS) that
occurred during 1992 to 1995 and was a birth cohort study of fluoride exposures and intake,
fluorosis and caries. The profiler did not put all assessments on the same profile as the data
were sufficiently different and provided different information. The reports are inter-related,
however, and will be grouped together. The reports include: Levy et al., 2003; Marshall et
al., 2004; Levy et al., 2006a and Levy et al., 2006b.
US/Iowa: 443 children who participated in the Iowa Fluoride Study; parents were recruited
in the hospital at their birth during 1992-1995.
PROFILER'S NOTE: The applicability of this study population for the general United
States is somewhat limited as the subjects and their families were mostly Caucasian, middle-
income and the majority of the parents had a college education.
None described
Parents of the children were recruited at birth. The children were given dental examinations
to identify fluorosis around age 9.
Children were part of the original Iowa Fluoride Study.
The study reported that the majority of the children were exposed to drinking water with
fluoride levels of 0.7-1.2 ppm.
PROFILER'S NOTE: Although the original IFS assessed fluoride levels from various
sources including, drinking water, foods, beverages, and the use of fluoride supplements and
dentifrices, this study only provided drinking water fluoride levels.
Analytical methods for evaluating the fluoride concentration in the drinking water were not
provided.
Children originally recruited into the Iowa Fluoride Study were examined at approximately 9
years old for dental fluorosis on early-erupting permanent teeth by two trained and calibrated
dentists using the Fluorosis Risk Index (Pendrys 1990). Twelve teeth were examined for
each subject: 4 mandibular incisors, 4 maxillary incisors and 4 first molars. After being dried
slightly with gauze, teeth were examined using a mouth mirror and exam light. Fluorosis was
differentiated from non-fluorosis opacities based on the criteria of Russell (Russell 1 96 1 and
Warren et al., 2001) and from "white spot' carious lesions based on color, texture,
demarcation and relationship to gingival margin.
The Fluorosis Risk Index (FRI) by Pendrys (1990) was used to assess fluorosis. The FRI was
developed to improve researchers' ability to relate the risk of fluorosis to the developmental
stage of the permanent dentition at the time of exposure to fluoride. The FRI assesses
fluorosis on four enamel zones classified according to the age at which fluoride enamel is
initiated. Ten early developing zones are defined as FRI-I zones (occlusal cusp areas of first
molars and incisal edges of 6 of the 8 incisors) while there are 24 FRI-II zones (that develop
and erupt later). Therefore, it has more potential to show accurate identification between age-
specific ingestion of fluoride and the development of permanent tooth enamel fluorosis.
186
January, 2008
-------
STATISTICAL
METHODS:
RESULTS:
Dental fluorosis
Using this index, three zones (incisal edge, incisal third and middle third) of facial surfaces
were assessed separately for these early erupting permanent teeth with the FRI scoring
criteria as follows: no fluorosis: FRI of 0; questionable fluorosis: FRI of 1, 50% or less of
zone with white striations; definitive fluorosis: FRI of 2, greater than 50% of zone with
white striations and severe fluorosis: FRI of 3, zone displays pitting, staining and/or
deformity. Cervical zones were excluded from the analysis because of incomplete eruption
of the teeth. Teeth that were unable to be scored were given a score of 9 and a tooth without
fluorosis was defined as having all zones with FRI of 0 (or 7).
PROFILER'S NOTE: The NRC (2006) states that the index by Pendrys (1990) was
specifically designed for use in case-control studies, of which there are very few. Levy et al.
(2006b) agree that this index was designed for case-control analytical studies of age-related
fluoride exposure risk factors for fluorosis with the purpose to maximize the contrast of the
identified case and control groups. The study authors also state that this method creates a
large number of scores being within the "questionable" category.
The only statistical analysis reported in the article was Kappa methods used to assess inter-
examiner reliability which ranged from 77% to 94% agreement.
Tables 1 and 2 are copied directly from Levy et al. (2006b). Tooth specific fluorosis
prevalence varied with the maxillary central incisor most affected and the mandibular
incisors least affected. With the three zones of the teeth, 40.6% overall had at least one tooth
with mild or more involved fluorosis, 30.2% had questionable fluorosis and 29. 1% had no
fluorosis. When only FRI zone I were considered, the percentages were 33.2%, 29.3% and
37.5%, for mild or more, questionable and no fluorosis, respectively. Using different
combinations of teeth, the prevalence using 3 zones were usually 1-9% points higher than
those estimated using FRI zone 1 only. Most fluorosis was mild with only 7 individuals
(1.6%) having FRI scores indicating severe fluorosis.
TABLE 1
Percentage of subjects with fluorosis by tooth*
Three zones; oodusal cabk/incisal edge, FW zone I only:
incisaJ third and middle third Incisal edge/ooclusal table
Non- Non-
Fluorosis fluorosis Fluorosis fluorosis
Tooth N cases* Questionable1 cases1 cases'* Questionable* cases'
3 443 17,2 22.8 60.0 16,7 223 60,9
7 443 21,2 26.0 52,8
8 443 275 23.0 49,4 23,7 20,1 56,2
9 443 26,6 25,7 47.6 22,6 21.9 55.5
10 443 16.9 27,3 55,8
14 443 14.0 19.9 66,1 13.8 19.9 66,4
19 443 11.1 17,8 71.1 9,3 16,5 74.3
23 443 4,1 9.7 86,2 4,1 8.6 87,4
24 443 2.9 8.1 88.9 1,8 5.9 92.3
25 443 2.5 8,4 89.2 1.4 6.1 92,6 .
26 443 3.2 9.0 87.8 3,2 7.9 88.9
30 443 11.3 17,6 71.1 9.3 ' 16.0 74,7
* Any tooth with any zone of score 9 (unable to score) on any of the non-cervical zones
of the 12 early-erupting permanent teeth was excluded from the table,
1 A tooth with fluorosis was defined as having a zone with FW score of 2 or 3.
'Questionable fluorosis was deined as having: a zone with FRI score oi 1, but no other zone
with a score of 2 or 3,
1 A tooth without fluorosis was defined as having all zones with FRI score of 0 (or 7),
187
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE
NOT FOUND IN NRC
(2006)
PROFILER'S DFG/1-07
REMARKS
PROFILER'S ESTIM.
NOEL/NOAEL
TABLE 2
Percentage of subjects with, fluorosis using different criteria
Number Percentage of subjects with fluorosis
Teeth of teeth Two or more zones with
Considered* required questionable fluorosis
(number of to show Sample Three on a tooth also
teeth) fluorosis size FRI zone P zones' considered as fluorosis
Incisors and 1 443 33.2 40,6 55.8
1st molars (12) 2 27.5 34.5 46.3
Incisors (3) 1 443 27.1 36.6 49.7
2 20.1 29.8 40,4
Maxillary central
incisors (2) 2 443 19.6 23.5 34.5
First Molars (4) 2 443 14,9 15,6 22,1
Maxillary central
incisors and 2 443 27.5 30,9 42.0
\* molars (6)
* Three' zones of each incisor and I* molar (12 teeth) must be scored to be included in the
table.
* Fiuorosis is defined as FRI score of 2 or 3,
PROFILER'S NOTE: The study did not provide details on the individual fluoride intakes as
related to the amount of fluorosis. For Table 1, there is no information identifying the tooth
number with the type of tooth, maxillary or mandibular incisors or the first molars (i.e. is
tooth number 3 an incisor or molar?). The data do not distinguish between a fluorosis score
of 2 or 3 which indicate definitive (2) or severe (3); the study only states that 1 individuals
had severe fluorosis which would be considered an adverse effect.
The FRI has advantages if used for analytical epidemiology studies for dental fluorosis;
however, the population prevalence varies depending on the index and case definition used.
The authors recommend that consideration be given to concurrent use of another index (i.e.
Dean's, TSIF) if prevalence estimates are an important study outcome.
In this study, the majority of the children drank water with the optimal fluoride level (0.7-1.2
ppm) and overall 34.5% had definitive fluorosis (FRI score of 2 or 3) on at least two teeth.
Russell, A.L. 1961. The differential diagnosis of fluoride and non-fluoride enamel opacities.
J Public Health Dent. 21:143-6.
Warren, J.J., S.M. Levy andMJ. Kanellis. 2001. Prevalence of dental fluorosis in the
primary dentition. Journal Public Health Dent. Vol. 61(2), p. 87-91.
The study interpreted data from the Iowa Fluoride Study and used the FRI to assess for
fluorosis, making the study not useful in determining a dose response. Quantitative data were
not included in regards to actual fluoride exposures and the number of children within each
fluorosis scoring group. The only quantitative data provided was the percentage of children
with "severe" fluorosis and this was only 7/443 (1.6%) indicating very few had an adverse
effect from fluoride. Data on statistical methods used and how fluoride levels in the water
were derived were incomplete. The study population was not a true reflection of the majority
of children in the United States as most were from families within the middle-income
bracket with parents having a 4-year college degree which tends to indicate better
preventative dental care was available.
Data are not suitable for development of a NOAEL for fluorosis.
188
January, 2008
-------
PROFILER'S ESTIM.
LOEL/ LOAEL
POTENTIAL
SUITABILITY FOR
DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
Data are not suitable for development of a LOAEL for fluorosis.
Not suitable (X), Poor (_), Medium (J, Strong (J
The use of the fluorosis risk index in identifying fluorosis in children.
189
January, 2008
-------
Mann, J., M. Tibi, H.D. Sgan-Cohen. 1987. Fluorosis and caries prevalence in a community
drinking above-optimal fluoridated water. Community Dent Oral Epidemiol 15:293-5.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental caries and fluorosis (permanent teeth)
Prevalence study of dental caries and fluorosis.
Israel administered Gaza Strip: adolescents (15-16 years old) residing in one small village.
All 182 (90 boys and 92 girls) of this age group participated in the study. All children
resided in the area since birth and had access to the same drinking water source (two local
wells) since birth No information was provided about written consent by the children's
parents.
None.
Birth to 15-16 years; the dates of the study conduct were not provided.
The concentration of fluoride in the well water was 5 ppm.
A questionnaire was administered to determine smoking, dietary patterns, drinking habits
and oral hygiene routines, but fluoride intake from these practices was not estimated.
Three samples of drinking water were analyzed for fluoride concentration (in ppm) on
different non-consecutive days using a combined activity electrode.
The objectives of the study were to assess the prevalence and severity of dental caries and
fluorosis in a community characterized by naturally above-optimal fluoridated water (5
ppm). The study population consisted of 182 (90 boys and 92 girls) adolescents (15-16 years
old) residing since birth in a small village in the Israeli administered Gaza Strip. All children
of this age group in the village participated in the study. Dental caries levels and the severity
of fluorosis were determined in all children. One study author conducted all the
examinations using artificial light, dental mirrors and sickle-shaped explorers. A
questionnaire was administered to determine smoking, dietary patterns, drinking habits and
oral hygiene routines, but fluoride intake from these practices was not estimated.
Dental fluorosis was determined according to Dean's index (see NRC, 2006, pages 88-89).
Dental caries were evaluated using the DMFS (decayed, missing or filled surfaces) index
according to the recommendations of WHO (1977). WHO criteria for classifying decayed
teeth were followed: only surfaces with detectable softened floor, undermined enamel,
softened wall or temporary fillings were recorded as DS (decayed untreated teeth).
Student's t-test, chi-square and analysis of variance were employed to analyze the results.
The level of significance was p<0.05.
Study results in Tables 1-4 are shown directly fromMann, 1987.
190
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/Date
REMARKS VAD/03-08-
07
T^Wf t Wn-ij** rv-,*«,g .-H.4 f'inJ '« jf. *IJ$, M*- FS OMFS" N Sfe'Ji
OS SM MS M> HS >i-» DW(:S SI »
Cemwi itWwiN OK' 0 Ift II Oil (KM n 0 W PW nc« •' <»i fiOO fi'-i'
Canlntt iHI 0('.» l» « IHH l'0»'< t'Hf u 'H :'• *
Is! iM«ti«l,tn 0 I1 « '*t i^ 'A< 8W Oft. t;>"«! i* !"* ,* "'
Jfitf pr«Htotor« fO! 4i k ••!.*,' «''** '!"-".' f"** l' " r»:'*i
:» «at*r* ??t ^*i* .!*• i "4 f-ffl «?S)« "'s: -1 U
?»4 t-.-i.n J"!U MO Ol»' I"*"1' 7«f!.S iKl' 2 it ' <*
low! 4 ?| ft 1ft DM JU Oni 00' SJfc > 1*
Tahle ? I,kv,».K«i nn^fsrtii' :i'J, ,' v>'j« U'A '.I1, IS PVf-'S', !>, f.t.n. i«% »',.,»
}'',v."frt« «i l^ SU Mi M> TS Sf> DMSS Nf»
Mi,?1 Si ?4: >M n« I« OCM UM 2K1 465
M«kT*ie «J 5 IS •< ^ O.W J ">l 000 !i fXi 442 * vi
fe\=ic *6 8X5 M* i S§ ?,« C02 'i h K ]7 tin,-
T«5*i 1*2 4 ?! <•!* 0?J Jtf l'(,! ,"IH « ?7 fN
F ^Mi*^ *l,i>!1, Si.' <iJiU) Hf BifKan1
l.»h,f ^ | "tulMMt k'xt'k h) W\
1 JUOHMIi H(5)S Otflt ToUll
Viij i' y- >i
Mrtte.i;,: U -i* fc1!
•>c«a- "^ * 4ft
K«4.i: */ T if?
i' «ttoni
liWr 4 Di^utitii«i t~» vif'iAu> icj'h fe) n.ti fim> ll* M,kl Mifc\TiiSf Scin-fi'
MJIM]!«;> 1«i rti.A'it-, I". r** i,l*.»
M^.Tdthti'ii'* Ki »rM'ljr» *V» J!*^ 3-i*.i
M* «r»!> ""-,••. niis| ir» !**, *' ", 20* »
M.>uijr,jj!'
-------
PROFILER'S ESTEM.
NOAEL
PROFILER'S ESTIM.
LOAEL
SUITABILITY FOR DOSE
RESPONSE MODELING
CRITICAL EFFECTS:
drinking water, was not estimated and could have explained the higher fluorosis levels in
boys. Analyses of drinking water for fluoride concentration were conducted three times, but
only one value (5 ppm) was reported. The range of levels should have been reported since
fluoride concentrations in water are known to fluctuate.
The incidence of severe fluorosis was about 25% for a drinking water fluoride concentration
of 5 ppm. Although other sources of fluoride exposure are possible, the incidence of severe
fluorosis is not inconsistent with the data of Dean (1942) for similar drinking water fluoride
levels.
The study design did not identify a no-fluorosis intake dose.
Study participants were exposed to only one concentration of fluoride in the water (5 ppm).
At this level, all the adolescents had caries and either mild, moderate or severe fluorosis.
Not suitable,O; Poor (X_); Medium (_); Strong (_)
Only a single exposure level was evaluated
Dental caries and fluorosis (permanent teeth)
192
January, 2008
-------
Mann, J., W. Mahmoud, M. Ernest, et al. 1990. Fluorosis and dental caries in 5-8 year-old
children in a 5 ppm fluoride area. Community Dent. Oral Epidemiol. 18:77-79.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
Dental fluorosis; dental caries
Survey
Gaza Strip: 152 children (72 boys and 80 girls), 6-8 yrs old, residing in a village in the
Gaza Strip. Comparisons were made with 16-18 yr old children from the same region
who had been evaluated in a previous study (Mann et al., 1987).
None
Six to 8 years (the date the examinations took place was not reported).
All children in the study population drank water from two local wells having fluoride
levels of 4.7, 5.3 and 5. 1 ppm (three samples taken on non-consecutive days).
Fluoride levels in well water were documented and the study authors noted that an earlier
study (Mann et al., 1985) had shown that tea consumed by the population in this region
had high levels of fluoride (levels and consumption patterns not reported), and that boys
drank more (unqualified amounts) tea than girls. Presumed (but unstated) assumption is
that monitored wells were the sole source of drinking water for examined children.
Combined F activity electrode (Orion Research of Cambridge, MA). Limits of detection
not provided by authors.
Dental fluorosis and caries in primary and permanent dentition were evaluated in a
population of children (72 boys and 80 girls, 6-8 yrs old), residing in a village in the
Gaza Strip who were exposed to fluoride in well water used as drinking water (4.7, 5.3
and 5. 1 ppm). Dean's index was used for scoring fluorosis and DMFS and defs for
scoring caries incidence. Data were analyzed statistically using Chi-square and
ANOVA.
Severity of fluorosis was monitored in primary and permanent dentition according to
Dean's index (Dean, 1942), with a range from normal enamel (score 0) to severe
fluorosis (score 4). Dental caries levels were also evaluated using the "DMFS and defs"
indices according to WHO (1987) recommendations. All clinical exams were conducted
by one examiner.
Chi-square was used to evaluate the prevalence of dental fluorosis in primary and
permanent dentition. DMFS and defs scores were evaluated with ANOVA. Level of
significance of p<0.05 was used for both methods. A Kappa statistic (of 0.83)
established a high level of inter-examiner reliability (92%).
For primary dentition, 45 of the 152 children (29.6%) had moderate fluorosis, but none
had severe fluorosis, and more females than males were fluorosis free (see Table 3,
copied directly from Mann et al., 1990):
193
January, 2008
-------
For the permanent dentition, 6 children had severe fluorosis and 55 had moderate
fluorosis (composite of moderate + severe was 40. 1%), and females and males had
similar levels of fluorosis. (see Table 1, copied directly from Mann et al., 1990):
i" 4' .
Other effects
The decay rate gradually increased in the permanent dentition with increasing fluorosis
severity, but the same pattern was not seen in the primary dentition.
A*. jVr-\. MS .LI i.i.-iii
STUDY AUTHORS'
CONCLUSIONS:
The study authors discuss several reasons for the observed lower severity of fluorosis in
the primary teeth: (1) the shorter duration of enamel formation and maturation of
primary teeth (and their much thinner enamel layer) allows a smaller amount of fluoride
to be deposited in the developing enamel (when compared to permanent teeth), (2)
primary teeth formed during the prenatal period do not receive the same level of fluoride
exposure from the mother as the permanent teeth from fluoride in drinking water, and (3)
higher exchange of fluoride during development of primary dentition as compared to
permanent dentition. The study authors note that the incidence of moderate to severe
fluorosis in the permanent teeth in the 16-18 yr olds [data from previous study (Mann et
al., 1985)] was higher in boys than in girls and may have been due in part to higher
levels of consumption (unqualified amounts) of tea by boys in this age group.
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
WHO. 1987. Oral Health Surveys; basic methods. World Health Organization, Geneva.
For description of Dean's Index of Fluorosis see Section 2, for definition of DMFS and
defs scores, see List of Acronyms.
PROFILER'S
DMO
Study not useful for defining NOAEL/LOAELS, but it does identify an adverse effect
194
January, 2008
-------
REMARKS 11/22/06
and
12/15/2006
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
level.
Not possible from the data presented (all children are reported to drink from the same 2
local wells).
The exposure level of 5. 1 ppm is considered an adverse effect level
incidence of moderate to severe fluorosis in the studied population;
cannot be identified.
due to the high
however, a LOAEL
Not suitable (J, Poor (_), Medium (X), Strong (J
Total fluoride intake due to consumption of drinking water and tea
is not documented.
Dental fluorosis and caries incidence
195
January, 2008
-------
Marshall, T.A., S.M. Levy, J.J. Warren, B. Broffitt, J.M. Eichenberger-Gilmore and P.J. Stumbo.
2004. Associations between intakes of fluoride from beverages during infancy and dental
fluorosis of primary teeth. Journal of American College of Nutrition. Vol. 23 (2): 108-116.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL
POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL
METHODS:
STUDY DESIGN
Dental fluorosis in primary dentition (and contributing variables).
Longitudinal cohort study
PROFILER'S NOTE: There are 4 articles (Levy et al, 2003; Marshall et al, 2004; Levy et al,
2006a and Levy et al., 2006b) that report on different findings but all use the same population
of children that were originally part of the Iowa Fluoride Study (IFS) during 1992 to 1995.
This IFS population was a birth cohort study of fluoride exposures and intake, fluorosis and
caries. The profiler did not place all assessments on the same profile as the data presented in
each report were sufficiently different.
US/Iowa: 677, 4.5-6.9 year old children (48.3% males and 5 1 .7% females) participating in the
Iowa Fluoride Study (IFS).
PROFILER'S NOTE: Marshall et al. (2004) questioned the applicability of this study
population for the general United States, as the subjects and their families were mostly
Caucasian, middle-income and the majority of the parents had a college education.
None described
Birth to 4. 5-6.9 years
Of the 690 children that received dental examinations at age 4.5-6.9 years old, only 677
adequately completed a questionnaire and could be used as the study population.
The study measured F in water, food, beverages and estimated F exposure from fluoride
supplements and dentifrices by the method of 3 -day food and beverage diaries every 3-4
months(from age 6 wks to 3 yrs) and then every 6 months.
As part of the IFS, non-municipal home and childcare water, filtered municipal water and
beverages were analyzed for fluoride. Concentrations of F from the public water systems were
obtained from the Iowa State Health Department. Fluoride levels of foods were assayed as
well. The method used to analyze for F was not described.
The Iowa Fluoride Study took place during 1992 to 1995; parents were first asked to
participate in the study during the hospital stay at the time of their child's birth. Parents of the
children were sent questionnaires on the children's diet and beverage intakes when the child
was 6 weeks old and 3, 6, 9, and 12 months old. After 12 months of age, each child was
examined every 4 months until the age of 3 years, and then every 6 months thereafter.
Children underwent dental examinations on the primary dentition between ages 4.5 and 6.9
years of age, conducted at The university of Iowa General Clinical Research Center.
Examinations were visual, conducted using a portable chair and exam light and performed by
two trained examiners. Dental fluorosis was determined using the Tooth Surface Index of
Fluorosis (Warren et al., 2001; Horowitz et al., 1984).
For the diet/beverage information, parents recorded their child's intake for 3 days and all
entries were compiled into a Food and Beverage Intake Table. A Nutrient Table was created
from nutrient data obtained from the U.S.D. A. A Fluoride Concentration table was also
compiled based on data from the water fluoride concentrations obtained and analysis of food in
the laboratory. A regional database then combined the data from the Food and Beverage
Intake Table, Nutrient Table and Fluoride Concentration Table to calculate the total daily
fluoride intake. Data were also collected on the usage of fluoride supplements and/or
196
January, 2008
-------
PARAMETERS
MONITORED:
STATISTICAL
METHODS:
RESULTS:
Variables as related to
presence of fluorosis
dentifrices.
The criteria used for identifying fluorosis in primary teeth were adapted from the Tooth
Surface Index of Fluorosis (Warren et al., 2001; Horowitz et al, 1984). Fluorosis was
distinguished from non-fluoride opacities based on differences in shape, demarcation and color
(Russell 1961).
PROFILER'S NOTE: Details of the TSIF index (Horowitz et al., 1984) are provided in
Section 2 of this report.
Statistical analysis was conducted using SAS (SAS, Version 8: Gary, N.C.). Subject
characteristics were categorized and presented as percentages. The Wilcoxon rank-sum test
compared distributions of beverage and fluoride intakes between subjects with and without
fluorosis. Statistical significance occurred at p<0.05. The categories making up the variables
were used to develop statistical models to predict fluorosis since not all of the relationships
between beverage and fluoride intakes and fluorosis were linear. Intakes were categorized into
three levels: none (nonconsumers), low and high intakes. Low and high intakes were defined
as below or above the median level of consumers. Each three-category variable was
represented by two indicator variables in fluorosis predication models using the LOGIST
procedure in SAS. Multiple logistic regression models were developed to predict fluorosis
status separately from beverage and fluoride intakes. Backward elimination was used to reduce
the number of variables and final models only included variables significant at p<0.05.
Estimated mean total fluoride intakes were 285, 396, 497, 539, 392 and 476 ug/day at 6 weeks,
3, 6, 9, 12 and 16 months, respectively. Estimated fluoride intakes from dentifrices or
supplements did not differ between subjects at any timepoint.
Part of Table 2 from Marshall et al. (2004) is copied directly below. Consumption of milk-
based formulas reconstituted from powder was associated with risk of fluorosis; quantities
consumed by subjects with fluorosis were higher and the fluoride concentration of water used
for reconstitution by subjects with fluorosis was higher. Multiple logistic regression models
developed to predict primary tooth fluorosis using categories of beverage intakes indicated that
high intakes of milk for 6 weeks through 16 months (p<0.05) were negatively associated with
fluorosis and high intakes of water used to reconstitute formula for 6 weeks through 16 months
(p<0.05) were positively associated . Multiple logistic regression models developed to predict
primary tooth fluorosis from fluoride intakes from various categories indicated that high
intakes of fluoride from water used to reconstitute formulas (pO.OOl) and from water as a
beverage (p<0.05) and any fluoride from supplements from 6 weeks through 16 months were
positively associated with fluorosis.
Beverage Fluorosis No fluorosis Wilcoxon p-value
Zr score
~j"u • -JIJIB 2 _'14~ « .~r|JNj:> 2">-t JJ>
*• lr^ It 3 5 J = I = *•*
"i-«. iniA )< (~ I i -JeJt i
L^-i 1 r>- IL * - < a^jll h * f * tj i^
Vl * i L|^. 0 XL* F»* ^" lO"* >r\ J". f "" ~ ~~2 1 ^
u~ 1 *=d 'SU ig _ -T ? t tr- JK"ar } d
j d-iu~nui» 2 ^4" i w i }2lf 3~"IC ~l.it
F-c^nrte." IT p^n' r ,. c t* J"f '44-bi -I'M I'l ". "'" n
F»^i_ntit«l i-i ...jy -c »ip.r« > " i ij" -..""i I f
Lc\ i_iB 'u4 ' C~ -2K'i On.3
] ' ! 4 ^ * £ i*vJ c i**d M*.-^stirt*' ^ fls*t*> „ s I 1 I ~2l C t"1
"" ta : i-\tj o- 4 '14i2i > c-i-Ji ; "~1 CM',
1 ii e J«^ Iri'*1 \Se" = ~ ]L « iu i J.C1"
a^al ^v ..e tL">oipc^ >-i > pv$ ^ u! 1 ni i — C S2! 1 1^
"j. •w-i cs -t "i^4 o" 'Jm ."4 11 i%2 I" .
PROFILER'S NOTE: Individual fluoride concentration levels were not provided in the study.
Also, the degree of fluorosis observed based on TSIF scores was not provided; fluorosis was
only stated as being present or absent.
197
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE
NOT FOUND IN NRC
(2006)
PROFILER'S Initials/date
REMARKS DFG/1-07
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM.
LOEL/ LOAEL
POTENTIAL
SUITABILITY FOR
DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
Infant beverages, particularly infant formulas prepared with fluoridated water, can increase the
risk of fluorosis in primary teeth.
Russell, A.L. 1961. The differential diagnosis of fluoride and nonfluoride enamel opacities.
Public Health Dent, 21:143-146.
Warren, J.J., S.M. Levy and MJ. Kanellis. 2001. Prevalence of dental fluorosis in primary
dentition. J. Public Health Dent, 61:87-91.
J.
The study cannot be used for dose-response modelling as no quantitative data was provided.
The study did not provide the degree of fluorosis identified (i.e. mild, moderate, severe) or the
amount of fluoride the children were exposed to in relationship to the fluorosis. Data on the
primary dentition for evidence of fluorosis were reported only as present or absent. The study
population was not a true reflection of the majority of children in the United States as most
were from families within the middle-income bracket with parents having a 4-year college
degree which tends to indicate better preventative dental care was available.
Some of the tabulated data may assist the relative source contribution analysis.
Data are unsuitable for development of a NOAEL for fluorosis.
Data are unsuitable for development of a LOAEL for fluorosis.
Not suitable (X), Poor ( ), Medium ( ), Strong ( )
Dental fluorosis in primary dentition (and contributing variables).
198
January, 2008
-------
Maupome, G., J.D. Shulman, D.C. Clark and S.M. Levy. 2003. Socio-demographic features
and fluoride technologies contributing to higher fluorosis scores in permanent teeth of
Canadian children. Caries Res 37:327-334.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
Dental fluorosis (maxillary anterior permanent teeth)
Canada/British Columbia: Prevalence study of dental fluorosis in male and female
children in grades 2, 3, 8 and 9 in two communities during 1996-1997. Data were from
British Columbia (BC) Fluoridation Cessation Study, a longitudinal evaluation of the
effects of removing fluoride from water supplies that were previously fluoridated in British
Columbia. The community of Comox/Courtenay/ Campbell River stopped fluoridating
water supplies in 1992 and was designated "FE" (fluoridation ended). The second
community, Kamloops continued fluoridation and was designated "SF" (still fluoridated).
All children included in the study were lifelong residents in these communities and had
permanent anterior teeth mineralized at the time water supplies were optimally fluoridated.
A total of 8,277 subjects were examined in 1996-1997, 49.6% fromFE and 50.4% from
SF. Of these, 49.8% were female; 50. 1% were in grades 2 and 3 and 49.9% in grades 8 and
9. Two dental examiners determined dental fluorosis in the children during 1996-1997
using the Thylstrup-Fejerskov Index (TFI; see NRC, 2006, pages 88-89 and Section 2 of
this report). Informed consent was obtained from parents and children, as approved by the
Ethical Review Board of the University of British Columbia.
None.
From birth until time of dental exam.
Children in grades 2,3,8 and 9 in two communities with differing F concentrations in
British Columbia, Canada.
Exposure measurements for the monthly fluoridated water levels (means ± SD) for 1985-
1992 in Comox/Courtenay and Campbell River, and for 1982-1993 in Kamloops were 0.92
±0.21, 0.88 ± 0.28 and 0.95 ± 0.27 mg F/L, respectively. Neither the analytical protocol
nor the water monitoring equipment used was reported in Maupome, et al (2003).
Based on data from questionnaires completed by parents, fluoride exposure histories
(supplements, rinses, toothpaste amount, tooth brushing frequency and tooth brushing
starting age) were developed.
None.
The present study used epidemiological data and fluoride and diet histories from follow-up
surveys and questionnaires previously described (Clark and Berkowitz, 1997a, b;
Maupome et al, 200 la, b) to examine levels of fluorosis among children from two
Canadian communities exposed to fluoride. Maxillary anterior permanent teeth of children
from these communities (grades 2, 3, 8 and 9 in 1996-97) were ranked with the TFI
scoring system (see NRC, 2006, pages 88-89 and Section 2 of this report).
Questionnaires completed by parents included information characterizing frequency of
dental attendance, use of bottled water or formula in the first year of life, frequency of
tooth brushing with home products containing fluoride, starting age of tooth brushing,
when self-brushing began, and the amount of toothpaste used during the first four years of
life (as a proxy measure for swallowing toothpaste and fluoride supplement use).
One of two trained dental examiners determined dental fluorosis for subject children. The
TFI intra-examiner reliability k coefficient 0.72 and inter-examiner reliability was 0.63 (on
199
January, 2008
-------
STATISTICAL METHODS:
RESULTS:
a per-child basis). Examiners viewed all teeth both with and without the use of the dental
light, and a score was assigned to each tooth. CDC standard procedures for
epidemiological dental examinations were followed (Summers et al, 1994).
Basic descriptive statistics were obtained and the relationships between TFI and the
independent variables using a multivariate regression model were assessed. The variables
obtained by questionnaires from the parents included age, gender and whether the children
lived in either SF or FE communities.
Generated variables derived from questionnaire data included: 1) a composite measure of
the socio-economic level (SES), frequency of dental attendance and parental education; 2)
pre- and post-eruptive exposure to fluorides through the use of fluoride supplements; 3) use
of bottled water or formula during the first year of life; 4) post-eruptive exposure to
fluorides from the frequency of tooth brushing with home care products containing
fluoride; and 5) estimation of the amount of toothpaste used in the first four years of life as
a proxy measure for swallowing toothpaste, together with tooth brushing frequency and
starting age. Since TFI scores were heavily skewed with a mode of zero, a multivariate
Poisson regression model was used. A fully saturated model was first fitted using TFI as a
dependent variable and the previously described covariates. Using the Wald X2 tests, the
least significant covariates were sequentially eliminated. All two-way interactions between
the covariates were tested and those that were significant (p<0.05) were added to the
model. If the two-way interactions were significant, three-way interactions were tested, and
those meeting the p<0.05 criterion were added to the model.
Study results are included in Tables 3 and 4 shown directly from Maupome et al (2003).
Table 3. Maximum TFI saves fur upper anterior teeth per child
TFI FEsite SFsite
frequency percent frequency peraent
(ch iMreu} (di ildr e 13 )
0 2.1)9 417 .1413 66. g
1 1.406 33.0 151 14.7
2 579 13,6 75* 14,8
3 126 3.0 153 3,0
4 19 0.0 W 0,0
4 13 0,0 [-! 0.0
6 3 0.0 1 0.0
200
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/
Table 4. Adjusted mean TFI score mid ditJerenoes ba^d cm Poison regression model, adjusting tor si^rtitiamt
coojounders, mcorjXTatirigsocio-deLTiog,raphii? data arid expomre to fluoride technologies
VarinWe Adjusted 95% Cf rafferenee
(least square) 2
fljectn TFI •*•
Sits- 8.02 (i.004fc
Com.Tv'Courtfnav CamphcllRivi-r(FE) 0.357 o,.Wi,r,42ft
kdmlonp,425
Mdl. 0.293 I'244,u352
Boitkdwkrutiml-uejronjli g.,05 0,0045
Ye, 0,267 i'2d4,t>35l
Ni.fLira.jtt.il 0,390 l> 342,0444
Child sutkd r'Ui"*lii£i£ *iEh to. ithfAt"^
1-2 iLdiM'tjjj.- 0.387 i' 323,0 4p2
., J-3,.ji 0,36 0.551,0
.. dJlti *rt.us 12,06 0.0005
J_\ ,M,,,| y,,, O.J71 0 M4 1'43?
i. dJtu 1 tu\ 10.48 0,0012
XUu'vMi^il IWL 0.235 0.176,0.315
Suppkmmtu > in M Vdi.itlit. 6,59 0.0102
Ni 0,285 i'22'-»,o35*i
Yes 0.366 mil.C'4H'
Fluoi idi mppknn ol t"i igut m. >
Itjih with Lv. mi'.^i^ 0.388 M ys o 4fr3
^H 4-hwpt-U 5.14 0.0234
«', uitKyuuith 12,40 0.0004
4-h units pttvurk 0.310 0,247,0.390
i-. iiifkciu.-iitK 0,74 0.3S84
Iiiliniuuith 0,280 0.227,0,345
Falta M'dur.iti.in 5.47 0,0194
Ci.llt'ei.'unin.isitv n 351 i'2'-*5,0419
High adi«..|-'tudc school n :w7 y 245. "3 5"
\& 3,77 0.0523
Ttndci l'i\fai\ f<2»5 ''245,0357
InaiHU.Idu 11.353 i<2W7,n4">l
REVIEWER'S NOTE: Residents of FE (fluoridation-ended in 1992; historical F
concentration
-------
REMARKS
Date
VAD/01-
02-07
PROFILER'S ESTEM.
NOAEL
PROFILER'S ESTEM.
LOAEL
SUITABILITY FOR DOSE
RESPONSE MODELING
CRITICAL EFFECTS:
other publications. The study was conducted in British Columbia, Canada, so the cohort
may not be representative of the U.S. general population. Recall biases could have
occurred since parents were asked to remember events that occurred many years prior to
completing the questionnaires and may have been confounded by memories of other
children. The examiners were different for each study site and were not blinded to site
fluoridation status; however, good k values suggested appropriate levels of examiner
consistency. Fluorosis results were generally limited to the low to mild category; 234 of
8,277 children (2.8%) had TFI scores of > 3.
This paper also contains information useful to analysis of relative source contribution.
Fluoridation status of study communities was in the range considered optimal for North
America North of Mexico (approximately 1 mg F/L). As a consequence, variation in TFI
scores between the study communities would not be expected to be large. It appears that
insufficient contrast in F concentrations between study communities was incorporated into
the study protocol. Further, fluoridation terminated in one study community in 1992, 4-5
years prior to dental exam. Thus, some confounding of collected dental fluorosis data is
likely.
The study design did not estimate a NOAEL.
The study design did not estimate a LOAEL.
Not suitable,(X); Poor (X); Medium (); Strong (_)
Dental fluorosis (maxillary anterior permanent teeth)
202
January, 2008
-------
Meneghim, M.C., Tagliaferrro, E.P.S., Tengan, C., Meneghim, Z.M.A.P., Pereira, A.C.,
Ambrosano, G.M.B., and Assaf, A.V. 2006. Trends in Caries Experience and Fluorosis
Prevalence in 11- to 12-Year Old Brazilian Children between 1991 and 2004. Oral Health
Prev Dent. 4(3): 193-198.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
Dental caries and fluorosis
Retrospective.
Brazil/Sao Paulo State/Iracemapolis: 236 male and female public school children, ages
11- and 12- years old; individuals were born in the town or lived there since the age of
two, did not use fixed orthodontic appliances, and did not present severe dental
hypoplasia. This was a follow-up in a series of surveys conducted in the same city
(Pereira et al., 2000; Kozlowski, 2001).
Comparisons were made with previous surveys of 1 1-12 year old children carried out in
Iracemapolis in 1991 (n=200); 1995 (n= 160); 1997 (n= 314); and 2000 (n= 244).
The study was conducted in 2004 on a group of children born in 1992 and 1993.
Fluoride was added to the water supply beginning in 1997 (7 years prior to the current
study). Fluoridated dental products became available in 1989. The study authors state
that children in all public schools performed tooth brushing under the supervision of
their teachers, and were instructed to brush twice per day; however, the report does not
indicate when the school-supervised program began.
Caries decline and fluorosis increase have been verified since 1991 (6 years prior to
water fluoridation).
Fluoride was added to the water at 0.7 ppm beginning in 1997. At the initiation of this
fluoridation program, the children in this study were already 4-5 yrs old. Fluoridated
dentifrices were commercially available since 1989, and there was a tooth-brushing
program at the public schools; consequently, children may have been using fluoridated
dentifrices since beginning school; however, the exact date when the school program
began was not indicated. The study authors state that no oral health program based on
fluoride therapy was available, and presumably this meant therapies such as non-
invasive application of fluoride in gel, solution, or varnish.
Dental caries and fluorosis prevalence were measured. No radiographs were taken
during the surveys.
Data on how fluoride concentrations in the water supply were measured were not
included in the study report. The fluoride concentration in the dentifrices also was not
included.
In the study, 236 schoolchildren (both genders), ages 1 1 to 12 years old, made up the
study population. All study participants were born in the town of Iracemapolis, Brazil
or lived there since the age of two, did not use fixed orthodontic appliances, and did not
present severe dental hypoplasia. The prevalence rates of dental caries and fluorosis in
the current study population (surveyed in 2004) was compared to prevalence rates from
previous surveys carried out in 1991 (n=200), 1995 (n=160), 1997 (n=314), and 2001
(n=244). The same protocol was followed in each survey.
Examinations: The decayed, missing, and filled permanent teeth (DMFT) index was
used for caries examination following the World Health Organization criteria, and the
Thylstrup-Fejerskov (T-F) index was used for fluorosis examination with the highest
score being registered for each child. Prior to the examination, each individual received
203
January, 2008
-------
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Caries
a toothbrush with fluoridated dentifrice and performed tooth-brushing supervised by a
dental hygienist. All schoolchildren were examined under natural light in an outdoor
setting, using a dental probe, buccal mirror, and air-drying. No radiographs were taken
during the surveys. For the DMFT index all permanent teeth were examined and for
the T-F index all the buccal surfaces of all permanent teeth that showed more than two-
thirds of erupted crown, and no fillings, were examined. The differential diagnosis
between very mild signs of dental fluorosis and nonfluorotic enamel opacities followed
the Russel's criteria.
Two dentists, aided by two recorders, performed the examinations in all the surveys.
One dentist examined dental caries and the other dental fluorosis. Duplicate
examinations were conducted in 10% of the sample to assess intra-examiner
consistency. Kappa values were greater then 0.81 for both examiners in all surveys.
Dental caries and fluorosis were measured using the mean number of decayed, missing,
and filled permanent teeth (DMFT) and Thylstrup-Fejerskov (T-F) indices,
respectively. No radiographs were taken during the surveys.
The variation of the DMFT index over time (1991 to 2004) was assessed by analysis of
regression at 1% significance level (pO.Ol). Comparison of fluorosis prevalence (T-
F>1) according to the year of survey was performed using the Chi-square test (p<0.05).
Table 1 was copied directly from Meneghim et al. (2006) and summarizes the results of
dental caries experience obtained in all surveys carried out between 1991 and 2004. In
2004, 50% of schoolchildren were caries-free. The mean DMFT was 1 .2, 82. 1% lower
than the results obtained in 1991 (DMFT=6.7). Seven years after fluoridation of the
water supply (2004), the DMFT was 58.6% lower (versus 1997 DMFT=2.9). In the
1991-1997 period, with no fluoride in the drinking water, the DMFT was lowered
56.7%. A significant decrease of DMFT over time for 12 year old schoolchildren
could be demonstrated by analysis of regression that showed a linear effect for DMFT
and year of survey (R2=0.92; figure 1, copied directly from Meneghim et al. (2006)).
Table 1 Mean DMFT and reduction {%) of cartes experience for schoolchildren aged 11 and
12 years in Iracemapolis, Brazil, according to year of survey
Year of survey Sample Mean DMFT
and Authors
1991 - Pereira et al, 2000 200 6,7
1995 - Pereira et al, 2000 ISO 3,9
199? - Pereira et al, 2000 314 2.9
2001 - KozIowsW, 2001 244 21
2004 - present study 236 1,2
J
» y = -0.4* ' t01,i
^ . R"' ' 9,V,if
31 4 *^v
O - ^^"^N,
! • """"^-^
'i* 1995 2000 H«6
Year
% Reduction in % Reduction in
relation to 1991 consecutive
41.8 41.8
58,7 25,7
68.7 27.6
82.1 42,9
Rg 1 DMFT vnnaticn for 11- to 12-year-oM schootehHdren in
f unct'ifn of t^P
PROFILER'S NOTE: The profiler agrees that there was a reduction in the mean DMFT
204
January, 2008
-------
Dental fluorosis
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date
REMARKS SJG/3/7/07
index over time from 1991 to 2004.
T
P
(I
1
a
K
P
ir
able 2 was copied directly from Meneghim et al. (2006) and summarizes the
revalence of fluorosis (T-F>1) from 1991-2004. Statistically significant increases
X0.01) were found between 1991 and 1997 as well as 1997 and 2004. In2004,
5.7% of children presented with fluorosis, T-F>1 . A total of 59.7%, 24.6%, 10.6%,
nd 5.1% of the children were scored as T-F=0, T-F=1, T-F=2, and T-F>3,
;spectively.
Table 2 Percentage of Individuals with TF>1 and increase (%) of fluorosis prevalence for
schoolchildren aged 11 and 12 years in Iracemapolis, Brazil, according to year of survey
Year of survey Sample Fiuorosis % Increase in % increase in
and authors prevalence relation to 1991 consecutive
(% TF>1) surveys
1991- Pereira etal, 2000 200 2.0
1995 - Perelta et al. 2000 160 4.4 120 120
1997 - Pereira et al, 2000 314 10.2 410 132
2001 - Kozlowski, 2001 244 12.7 535 25
2004 - present study 236 15.7 685 24
ROFILER'S NOTE: The profiler agrees that the prevalence of fluorosis (T-F>1)
icreased over time.
A significant decrease in dental caries and a significant increase in dental fluorosis
prevalence could be verified from five epidemiological studies carried out from 1 99 1 to
2004 in Iracemapolis, Brazil. The expansion of preventative programs at schools (date
of initiation not reported), the presence of fluoride in the water supply (since 1997), and
fluoridated dentifrices (available since 1989) are recognized as the main factors for
caries decline in Brazil.
Most of the individuals with fluorosis presented with T-F scores of 1 or 2 (T-F>3
=5. 1%) which does not affect aesthetics or function. Water fluoridation can not be
claimed as the responsible factor in the increase of fluorosis prevalence observed from
1997 to 2004. The significant increase in fluorosis prevalence is possibly due to the
inappropriate use of fluoridated dentifrices by young children.
Kozlowski, F.C. (2001). Relacao entre o fator socioeconomico e a prevalencia e
severidade de fluorose e carie dentaria. [Dissertacao]. Piracicaba: Universidade
Estadual de Campinas, Faculdade de Odontologia de Piracicaba.
Pereira A.C., Cunha F.L., Meneghim M.C., Werner C.W. (2000). Dental caries and
fluorosis prevalence study in a nonfluoridated Brazilian community: trend analysis
and toothpaste association. ASDC J Dent Child. 67, 132-135.
Russell, A.L. (1961). The differential diagnosis of fluoride and non-fluoride enamel
opacities. Journal of Public Health Dentistry 21, 143-146.
World Health Organization (1987). Oral Health Surveys: Basic Methods. 3rd ed.
WHO, Geneva.
World Health Organization (1997). Oral Health Surveys: Basic Methods. 4th ed.
WHO, Geneva.
The study was well-conducted and had adequate study design. However, the study was
not designed for development of a dose-response to fluoride as the emphasis was on
monitoring dental caries and fluorosis in children before and after the addition of
fluoride in the water supply (1997). Based on the study design, the observed decrease
in caries and increase in fluorosis from 1 99 1 to 2004 can not be attributed to any one
cause. A limitation of the article is that a description of the preventative program in
schools (i.e., brushing under teacher supervision) was not described fully, including
205
January, 2008
-------
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
when it was initiated.
The increase in fluorosis prevalence can not be attributed to water fluoridation because:
1) A significant increase in fluorosis was observed from 1991 to 1997, prior to the
addition of fluoride to the water in 1997 (other sources of fluoride exposure, besides
fluoridated denitricies, were not mentioned). 2) The percent increase was higher before
1997 (120% between 1991 and 1995 and 132% between 1995 and 1997) than afterl997
(25% between 1997 and 2001 and 24% between 2001 and 2004); and 3) Examinations
were conducted on children after the critical period for developing fluorosis. Children
examined in 2004 were 4-5 years old in 1997 and the critical period for developing
manifest fluorosis in the upper central incisors is from 2-3 years old
PROFILER'S NOTE: The study authors state that the critical period for developing
manifest fluorosis in the upper central incisors is from 2-3 years old, but upon further
inspection of the reference (Ishii and Suckling, J Dent Res 1991; 70: 952-956), this is a
questionable time frame. Ishii and Suckling concluded that children were "at-risk" for
fluorosis development if exposure to high fluoride in the water (7.8 ppm) started after
1 1 months and ended before 7 years of age, and this time frame was dependent on tooth
type (i.e., pre-molars, upper central incisors, molars). Although estimates for critical
periods were reported, limitations in their study preclude conclusions for "at-risk"
periods of susceptibility to fluorosis from being used without caution. Therefore, the
children examined in the current study still may have been vulnerable to fluorosis in
1997 when water fluoridation was initiated (they were less than 7 years old).
Study design was not suitable for development of a NO AEL for caries or fluorosis.
Study design was not suitable for development of a LOAEL for caries or fluorosis.
Not suitable (X), Poor ( ), Medium ( ), Strong ( )
While the study was well-conducted, the study design was not conducive to provide
data for a dose-response. The study only indicated an increased incidence of dental
caries and decreased prevalence of fluorosis in the public school population of 1 1- to
12- year old children in Iracemapolis, Brazil from 1991 to 2004. The study did not
address any issues of plaque or gingivitis.
Prevalence of dental caries and fluorosis
206
January, 2008
-------
Olsson, B. 1979. Dental findings in high-fluoride areas in Ethiopia. Community Dent. Oral
Epidemiol, 7:51-56.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
Dental fluorosis, dental caries and gingivitis in schoolchildren
Cross-sectional survey
Ethiopia/Rift Valley. 478 children aged 6-7 years old and 13-14 years old from either Wonji,
a sugar plantation area in the Shoa province or Awassa, the capital of the Sidamo province,
were used in the study. 60 boys and 60 girls in each age group for each area were randomly
selected for the study; however, data from 2 children were lost.
None
The exact duration of exposure to the drinking water was not stated, however, the older
children were questioned on the place of their birth, thus assuming the children were exposed
to drinking water in the area since birth. An exact date on when the study took place was not
provided.
PROFILER'S NOTE: There was no information regarding whether or not the 6-7 year old
group were asked the place of their birth; the profiler assumes Olsson (1979) thought they
would have been native to the area.
The schoolchildren were from either Wonji or Awassa which are in the Rift Valley area of
Ethiopia. Fluoride levels in Wonji, taken from six wells, ranged from 6.0 ppm to 17 ppm
(mean of 12.4 ppm). In Awassa, fluoride levels taken from seven wells ranged from 1 .2 ppm
to 7.4 ppm (mean of 3.5 ppm). Dates for well water sample collections were not reported.
The children in the 13-14 year old group were asked whether they were born in the area to
determine if exposure to the drinking water had been since birth. The study did not state if the
younger children were asked this information. No information was included in the study on
other possible sources of fluoride besides the drinking water.
Water samples were analyzed for fluoride content from six wells in the Wonji area and seven
wells in the Awassa area. The method of fluoride analysis was according to that of Frant and
Ross (1966).
478 schoolchildren aged 6-7 years old and 13-14 years old from either Wonji, a sugar
plantation area in the Shoa province; or Awassa, the capital of the Sidamo province were
evaluated. Subjects were selected from schools in the community from randomly chosen
classes until 60 boys and 60 girls in each age group were identified. Water samples from
wells in both areas were collected and analyzed for fluoride. Olsson served as the single
dental examiner for the subjects; thorough dental examinations were performed under natural
light using plane mouth mirrors and explorers for dental fluorosis, oral hygiene assessment,
periodontal condition, and dental caries.
Dental fluorosis was scored according to the criteria of Dean (1934), as modified by Moller
(1965). The severe score was reserved forteeth with extensive loss of enamel while teeth
with some confluent pits only were scored as moderate. Each individual child was assigned a
fluorosis rating according to the two teeth with the highest fluorosis score. Also, a community
index of dental fluorosis (FC1 ) (Dean 1942) was determined for each area using the formula
below. Frequency was the actual number of cases in each dental fluorosis score category
times the weights for each score, which were as follows: very mild, 1; mild, 2; moderate, 3;
and severe, 4. The total numerator was divided by the total number of children examined. A
similar severity index of fluorosis was determined for tooth groups (F^ ) (Moller et al. 1970).
F^ = frequency x weight
No. of individuals
207
January, 2008
-------
Oral hygiene was assessed by the Simplified Oral Hygiene Index (OHI-S) as presented in
Greene and Vermillion (1964). Periodontal condition was assessed as sound ("sound" was
not defined) presence of gingivitis only, or presence of destructive periodontal disease,
following the criteria of WHO for field surveys (1971).
Dental caries was diagnosed for the tooth surface only when definite cavitation existed "with
sticking on probing" occurred. Assessments were made on the number of decayed and filled
primary teeth (dft) and the number of decayed, missing and filled permanent teeth (DMFT).
Duplicate examinations were performed on 18 of the 13-14 year olds. In duplicate
examinations, caries assessments were reproduced in all instances but one and the fluorosis
scoring was consistent in 89% of the examinations. The different scorings occurred mostly in
the very mild and mild groups.
PROFILER'S NOTE: Olsson (1979) rated the fluorosis according to Dean (1934). Dean's
Index of Fluorosis is described in Section 2 of this report. The difference between the two
indexes was that the 1934 fluorosis scoring by Dean (1934) included a category of
"moderately severe and severe," and the modified 1942 Dean index combined these
categories into one, calling it "severe" (NRC 2006). The profiler is unsure what is meant by
".. .as modified by Moller (1965)," but has ordered an open-literature paper by Moller et al
(1970) to investigate the issue.
STATISTICAL METHODS:
No information characterizing the type of statistical analysis used was included in the study
report. However, Olsson stated that statistical analysis was performed on the score
frequencies.
RESULTS:
Dental fluorosis
Tables 1 and 2 are directly copied from Olsson (1979) and provide data on the fluorosis
scoring and incidence. Olsson (1979) stated that no gender-related differences were observed
in any of the parameters examined. In Table 1, in both cities, moderate and severe fluorosis
was more prevalent in the permanent dentition compared to the primary dentition, and both
communities of Wonji and Awassa had a similar number of participants with moderate and
severe fluorosis scores. Tables (not copied below) were also provided indicating the tooth
severity index and identified the second molars as the teeth most affected by dental fluorosis
in both the primary and permanent teeth in the children born in Wonji and Awassa.
Table 1. Frequency di«;r;bmior» of 239 children, aged <]. '
year;, from Vtonji '12,4 par, ar.d Awas
-------
Dental caries
Other findings
not native to the area. All of the 13-14 year old children born in Wonji and Awassa had dental
fluorosis with most having a moderate score. For those not born in the areas, 84% had dental
fluorosis with the majority scoring as very mild.
fable 2, Fieq.iearv dibtiibddou of 239 children,, aged ]>-
14 years, horn Wonji (12.4 par.s/10** F j and Aivasw. ,'*s~
pans/ID6 F ) acconLng' to decree of dental fluorosis am'
(ouhi.uum index of dfitcu Luorosis J7,.^ in the pvn.w-
ncut deniuicn In children boai and not born in the a.rtu
Children Chihiieti
Denial Iluorosis «_, *,. •
•i% » iNOl T> * i^**
crnrp Born m JJom m
Wonji bom111 Aw** borr'm
\Vonjt Awayu
None - 5 - ',1
Very mi id - 12 4 t(>
Mild 6 6 5 12
Moderate 85 9 43 5
Severe 16 - -
No, of persons 87 32 52 SB
Pci 3.11 1.59 2.75 I. Hi
PROFILER'S NOTE: A confounder was identified for the data from Wonji as Olsson (1979)
stated at the end of the paper that some wells located in Wonji had a fluoride concentration of
2 ppm, and defluoridated water was distributed to the living quarters of the management of
the sugar factory and also to some of the plantation villages. Data are unclear regarding
whether this occurred during the duration of the study or if all or any children included in the
study were either living in the management quarters or within the plantation villages receiving
the defluoridated water. These data must be considered when applying the study to use for a
dose-response. Children in Wonji did have severe fluorosis but if they were exposed to some
water with lower fluoride concentrations, the number of children with reported scores of
moderate or severe may be lower than if they had been exposed only to the water in Wonji
with the reported fluoride levels.
Table 5 was copied directly from Olsson (1979) and shows how the fluorosis scores correlate
with the incidence of caries in both the 6-7 year old children's primary teeth and the 13-14
year old children's permanent teeth with further breakdown into first and second molars.
Caries was diagnosed in 24% of all teeth with severe fluorosis.
J 1^ V * f!V*l
PniiM' ff t ]'<•.' ni vi It., i I'l'i iii! i-* Vijnd rfi'.ip
'"'" '' '"'"'' x „ Jen.'.! N| >HUd ^^ (| [Irr,. r' ^n (j IVfi,Pt
u-i," v/' ' "ti"' Th "-' 7-1' "•'''' Tj'
Now '>>>' J "I"* ''* ~ Jfi '
Vf ' i r . li'i , " '! ' 5i " Jl) S
\4 ^ i JM j if ^ i ?'^j r> n^ t(
\'\,"t.,i, 4 t s ,/u i rs i n',
Srvi-r, .'1 .4 u' Ti '7 H C
PROFILER' S NOTE: In both primary and permanent teeth, the number of decayed teeth
increased when the dental fluorosis score was severe.
Olsson (1979) also reported that gingivitis was observed in 99% of the 6-7 year old children
and 95% of the 13-14 year old children. The OHI-S was 1.99 ± 0.75 for the 6-7 year olds and
1.89 ±0.67 for the 13-14 year olds.
209
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S DFG
REMARKS 11/30/2006
and
12/14/2006
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM.
LOEL/LOAEL
POTENTIAL
SUITABILITY FOR DOSE-
RESPONSE MODELING:
CRITICAL EFFECT(S):
From the study, it was concluded that severely disfiguring fluorosis was very common in the
surveyed areas, causing an increase in both caries and gingivitis in addition to the obvious
aesthetic disadvantages. Olsson (1979) stated that his results were probably lower than other
comparable studies because he did not score questionable fluorosis and that only teeth with
extensive loss of enamel were given the severe score. He also noted that high water
consumption in the dry season, frequent tea drinking and malnutrition could have attributed to
the fluorosis present. Olsson (1979) stated that data indicated conclusive evidence of placenta!
transfer of fluoride in high-fluoride areas due to moderate fluorosis in the enamel of primary
incisors.
Frant, S.M. and J.W. Ross, Jr. 1966. Electrode for sensing fluoride ion activity in solution.
Science, 154: 1553-1554.
Greene, J.C. and J.R. Vermillion. 1964. The simplified oral hygiene index. J. Am. Dent.
Assoc., 68:25-31.
Moller, I.J. 1965. Dental fluorose og caries. Thesis. Rhodes, Copenhagen.
Moller, I.J. , J.J. Pindborg, I Gedalia and B. Roed-Petersen. 1970. The prevalence of dental
fluorosis in the people of Uganda. Arch. Oral. Biol. 15: 213-225.
WHO (World Health Organization). 1971. Oral health surveys: basic methods. Geneva.
Moderate and severe fluorosis was observed in this study mostly in the permanent dentition of
children native to the area. While Olsson (1979) recognizes that other factors could be
contributing to the high scores besides the fluoride in the drinking water, i.e. tea drinking,
malnutrition, those factors were not addressed further in this study. Some confounding
information was identified at the end of the report. Olsson (1979) stated that some areas
around Wonji had wells with lower amounts of fluoride (2 ppm) and that defluoridated water
had been distributed for several years to some residential areas and villages (not separately
analyzed) in the study area. Because it is not clear if the children in the study were included in
this population given the lower fluoride water for several years, this creates a possible
confounder that could make the reported data for the community of Wonji actually lower than
if the children had been exposed to the measured quantities of fluoride continuously.
Adequate data concerning the statistical analysis used, if any, were lacking.
Data are not suitable for development of a NOAEL for fluorosis.
Data are not suitable for development of a LOAEL for fluorosis.
Not suitable (J, Poor O, Medium (X), Strong (X )
While the study does provide adequate study numbers and follows the standard indexes of
measurement for fluorosis, data gaps (i.e. other sources of fluoride, lack of statistical analysis)
and confounding information (the possibility of some children being introduced to
defluoridated water in Wonji for an undetermined duration) makes this study of medium use
for dose-response. There were adequate data to correlate a severe fluorosis score with a high
level of decay in teeth when children were exposed to high levels of fluoride for some period
of time. Data for Awassa was not affected by the confounding data.
Dental fluorosis, dental caries and gingivitis in schoolchildren
210
January, 2008
-------
Retief et al., 1979. Relationships among fluoride concentrations in enamel, degree of fluorosis and
caries incidence in a community residing in a high fluoride area. Journal of Oral Pathology
8: 224-236.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Caries
Dental fluorosis and dental caries
Cross-sectional study of fluorosis and caries incidence and fluoride levels in teeth
enamel
South Africa/Kenhardt: 85 children, 14-16 yrs old (37 boys and 48 girls) of mixed racial
background (Caucasian, Negroid and Malayan). The town drinking water is obtained
from boreholes. No further information on the study population was reported. Rainfall
in the study area was reported to be 12.5-25 cm per year; the mean maximum annual
temperature was not reported.
None
Not stated
The fluoride content of the water supply was 3.2 ppm
Fluoride exposures from non-drinking water sources were not evaluated.
The fluoride content in the teeth enamel was determined with a Model 96-09 Orion
combination fluoride electrode coupled to an Orion Model 80 1 A pH/mv meter.
The relationship between fluoride concentrations in teeth enamel and dental fluorosis
and dental caries was investigated in a South African children, ages 14 -16, who lived in
an area with approximately 3.2 ppm fluoride in the drinking water. The children were
examined by a dental surgeon and the caries incidence (DMFT), and the degree of
fluorosis, measured by Dean's Index of Fluorosis (see Section 2) were assessed. The
fluoride content of the children's teeth enamel was analyzed and the fluoride
concentrations were corrected to a uniform depth of 10 um. The association between the
degree of dental caries, dental fluorosis, and fluoride content in the teeth enamel was
assessed.
The DMFT index and the percentage of children free from caries was used to measure
the incidence of dental caries and Dean's Index of Fluorosis was used to evaluate the
grade of dental fluorosis
The data were transformed from mass fluoride in picograms to log mass fluoride and
from mass enamel in micrograms to log mass enamel. Preliminary t-tests were carried
out to determine the significance level of the differences between log mass fluoride and
log mass enamel for the left and right incisors of the children. A covariance analysis was
run to test the differences between sexes for log mass fluoride using log mass enamel as
the covariate. The adjusted log mass fluoride mean values were obtained by correcting
the unadjusted log mass fluoride values to a standardized depth of 10 um. In order to
determine the association between the degree of dental caries, the degree of dental
fluorosis, and log mass fluoride, correlations between the various pairs of variables were
calculated adjusting for the sex of the children.
See Table 2, copied directed from Retief et al. ( 1 979), for the incidence of dental caries
(DMFT) and the percentage of children caries free:
211
January, 2008
-------
Dental fluorosis
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date
REMARKS SBG
3/28/07
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
Table 2. Results of clinical and laboratory evaluations
Male Female
37 48
DMFT(±S.E.) 3.35*0.42 2.08+0.33
% Caries free 16.2 . 35.4
DEGF(±S.E.) 2.86+0.06 2.71+0.09
% Severe fluorosis 86,5 77.1
Biopjy depth (+S.E.) um 10.08±0.20* 9.43+0.32*
Mean unadjusted
F(+S.E.)ppm 5078+261* 5040±319*
* Mean and S.E. of the average values for left and right maxillary central incisors of each individual.
DMFT vs. log mass fluoride, r = .415, PO.0001
The degree of dental fluorosis (DEGF) in the children is shown below in Table 2 copied
directly from Relief et al. (1979:
Table 2. Results of clinical and laboratory evaluations
Male Female
37 48
DMFT(±S.E.) 3.35+0.42 2.08+0.33
% Caries free 16.2 . 35.4
DEGF(+S.E.) 2.86+0.06 2.71+0.09
% Severe fluorosis 86,5 77.1
Biopjy depth (+S.E.) um 10.08±0.20* 9.43+0.32*
Mean unadjusted
F(+S.E.)ppm 5078+261* 5040±319*
* Mean and S.E. of the average values for left and right maxillary central incisors of each individual.
DEGF vs. log mass fluoride, r = 0.389, PO.005
DMFT vs. DEGF, r = 0.25 1, PO.02
The exposure to drinking water with a high fluoride concentration (3.2 ppm) resulted in
86.5% of the males and 77. 1% of the females showing severe fluorosis associated with
marked pitting of enamel. A highly significant positive correlation between the degree
of dental fluorosis and the fluoride enamel levels was seen, as well as a significant
association between the incidence of dental caries and fluoride concentrations in the
teeth enamel.
None
This study was in South Africa and there was no control group with exposure to lower
levels of fluoride in the drinking water.
Confounding factors which may have contributed to total fluoride intake, such as dietary
habits, were not addressed and only one exposure level was considered. Nevertheless,
the results might be useful when compared with data from other studies.
A NOAEL for severe dental fluorosis was not identified in the study
Based on the results from Table 2, the fluoride concentration of 3.2 ppm was associated
with a high incidence of severe dental fluorosis; therefore, the lowest effect level would
be expected to be less than 3.2 ppm.
Not suitable ( ), Poor (x ), Medium ( ), Strong ( )
212
January, 2008
-------
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
This study itself is poorly suited for dose-response
was evaluated.
modelling because only one dose
Dental caries, dental fluorosis
213
January, 2008
-------
Rozier, R.G. and G.G. Dudney. 1981. Dental fluorosis in children exposed to multiple sources of
fluoride: Implications for school fluoridation programs. Public Health Rep 96(8):542-48.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
Dental fluorosis (permanent teeth)
Prevalence study of dental fluorosis.
North Carolina (rural area, exact location not stated): The study population originally
included 307 children, ages 5 to 15, in grades kindergarten through 8. Residence and
water histories were obtained through serf-reporting and verification by the school
principal.
None.
Children were continuous residents of the community so their exposure period was from
birth to the age at which study examination took place (5-15 years old). The date(s) the
study was conducted were not provided.
The school used well water with a natural fluoride concentration of 4.5 ppm (5.6 times
the recommended water fluoride level of 0.8 ppm for community fluoridation in that
area). The final study population was divided into four groups based on the fluoride
concentration of the home water supplies: Group 1, 0.00-0.24 ppm (average 0.18 ppm);
Group 2, 0.25-1.99 ppm (average 0.87 ppm); Group 3, 2.00-3.99 ppm (average 3.13
ppm); and Group 4, 4.00-6.50 ppm (average 4.82 ppm).
Samples of home water supplies were collected from 292 of the students.
The fluoride concentration of the water samples was determined by the electrode method
(Harwood, 1969).
The objective of the study was to determine if continuous, lifetime use of home drinking
water naturally fluoridated to optimum levels combined with the use of school water
having 4.5 ppm natural fluoride, beginning at school age, causes objectionable levels of
dental fluorosis in school-age children in rural southeastern North Carolina. The study
population of children, ages 5 to 15, in grades kindergarten through 8 was examined for
the presence or absence of dental fluorosis according to the definitions and examination
criteria of Dean (see NRC 2006, pages 88-89). All examinations were performed in a
designated room in the school; portable dental chairs and lights, a No. 23 explorer, and a
plane surface mirror were used.
Color transparencies were made at the time of the examination. A sample of these
representing the spectrum of fluorosis severity were later reviewed and scored by an
independent investigator. Residence histories and water supplies were obtained through
serf-reporting and verification by the school principal. Only data from children who
were continuous residents of the community and who returned a water sample were
included in the study analyses (22 1 of 295). The study population was divided into four
groups based fluoride concentration in their home water supplies. For each group, a
qualitative and quantitative Community Index of Dental Fluorosis (Fci) was calculated
according to methods suggested by Dean (1935). To measure the effects of home and
school fluoride exposure, only students with at least one erupted premolar or second
molar were included in the analysis, resulting in a final sample of 120 students.
Dental fluorosis was evaluated using a modified Dean's Index in which two scores were
determined on each student, one representing fluorosis prevalence in early erupting
permanent teeth and one in late erupting permanent teeth. Each score was based on the
most severely affected tooth in each group of teeth. At least one tooth of the group had
to be present for a score to be assigned.
214
January, 2008
-------
STATISTICAL METHODS:
RESULTS:
STUDY AUTHORS'
CONCLUSIONS:
A quantitative and qualitative Community Index of Dental Fluorosis (Fci) was
calculated for each of the exposure groups. To determine the quantitative index, a weight
was assigned to each of six classifications of fluorosis given at the time of the
examination, and the weighted scores for all individuals within each group were
averaged. For the qualitative index, seven descriptive terms - negative, borderline,
slight, medium, rather marked, marked, and very marked - were derived from the
percentage distribution of the six classifications of fluorosis.
None were described in the study report.
The study results are provided in Tables 1-3 taken directly fromRozier, 1981.
Tiifcln 1, Comparison ol pf'rMaejrapha and c-mieal ej«am>nfl-
lions for the prHM>M»r;« or absent-*- of ite^tal Hoofe-sis fo» 63
locih Svara-B ul 41 school children
KIL 't. 3fjf«ft - » - - "tnt
N'crTa' 1T 1 12
Mm5roSi3 , , . 3 40 5f
Total* 14 49 $3
f'HCfsds corinET1* ot ?hf? T h-3Hii" ^<$*!pr &^ppi'*.is
*.,«.„., r«r*,*r
fg* frartf &timt*t i 1 vy *' ^ji« "*> Wh**(- Pin- h ef ffl i*ji.* ff Wiw >>*»r
000-B24 t! 18 3U 11 Btt 3 *0 3 10
@2Mfl'} 0 B7 F«i 14 *>fl 7 5S 3 ~S * 4
ieo-399 aij Jh « *'J e 17 14 *o a a 3 s i 3
•tOO-tm 4K 3>~> I ,1 t 13 1C S3 b 20 3 K
Tolal 2,^4 1JJ SJ 44 X t," 30 "5 10 i t- S 1 1
Table 3. Fsuofftsis ^de.«; fp*1 childr*fi *«5f> at east 1
-------
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/Date
VAD/03-13-
07
PROFILER'S ESTEM. NOAEL
PROFILER'S ESTIM. LOAEL
SUITABILITY FOR DOSE
RESPONSE MODELING
CRITICAL EFFECTS:
impairs appearance.
Harwood, J.E. 1969. The use of an ion-selective electrode for routine fluoride analyses
on water samples. Water Res 3 :273-280.
Dean, H.T. 1942. The investigation of physiological effects by the epidemiological
method. In Fluoride and dental health, edited by F.R. Moulton. American
Association for the Advancement of Science, Washington, D.C., No. 19, pp. 23-
31.
In general, there was a good dose-response relationship between increased frequency
and severity of fluorosis and increasing concentrations of fluoride in the water supplies
(Table 2). However, study flaws were noted. The demographics of the study population
were not described; therefore, no determination can be made about whether it was
representative of the general U.S. population. Exposure to other sources of fluoride, in
addition to drinking water, was not considered. The number of each gender of the study
population was not provided. Four examiners evaluated the children for fluorosis, but no
measure of inter- or intra-examiner consistency was provided.
The LOAEL for severe fluorosis occurred at or above 2.0 ppm F in home drinking water
(only 1 of 65 children); however, the overall exposure was greater because the fluoride
level in the school drinking water was 4.5 ppm. Further, the data were not analyzed
statistically to determine if the occurrence of severe fluorosis was significant. The fact
that severe fluorosis was not seen at 4.0-6.5 ppm F in their home drinking water raises
questions about the methodology used. If the children were examined after all
permanent teeth had erupted, the overall incidence rate is likely to have been different.
The study design did not identify a no-fluorosis intake dose.
The LOAEL for fluorosis in children aged 5-15 years was 0.00-0.24 ppm. Children at
the lowest range of fluoride concentration in their home water supplies had mild
fluorosis.
Not suitable,(_); Poor (X); Medium (_); Strong (_)
Dental fluorosis (permanent teeth)
216
January, 2008
-------
Ruan, J.P., Z.Q. Yang, Z.L. Wang, A., Astrem, A., Bardsen, A., and Bjorvatn, K.. 2005a. Dental
fluorosis and dental caries in permanent teeth: rural schoolchildren in high-fluoride areas in
the Shaanxi province, China. Acta Odont. Scand. 63: 258-265.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION
STUDIED:
CONTROL
POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
Dental fluorosis and dental caries (DMFS = decayed, missing, filled tooth surfaces) in
permanent teeth
Case control, retrospective
China/ Shaanxi Province: 477 schoolchildren from two rural areas (Bao Ji County and Jing
Bian County). The province is in northwestern China between 3 1-39°N and 105-1 10°E. The
children were age 12-13, approximately half male and half female, and attended 13 village
primary schools in communities with comparable socioeconomic standards. The drinking water
source was groundwater obtained from 28 deep (20-30 m) wells, and was unchanged over the
last>13 years.
For most of the comparisons made there was no control population per se, rather, children were
subdivided into five fluoride exposure groups that were compared to one another. For
evaluation of the relationship between the prevalence of dental fluorosis (TF score of >3) and
the associated fluorosis risk factor, the following groups were used as the reference groups for
the analyses: those that did not use (local) clay pots in which to store drinking water compared
to those that did, fluoride group A (0.3-0.5 mg F/L) compared to other concentrations, males
compared to females, and 12-year olds compared to 13 year olds (see Table V of Ruan et al.,
2005a).
Lifetime (to ages 12-13).
Schoolchildren aged 12-13 from two rural areas in Shaanxi Province, China, who obtained their
lifetime water supply from one of 28 wells. The children were subdivided into five fluoride
exposure groups (A, B, C, D, E) based on their well water fluoride concentration, as shown in
Table I. The distribution of the children in the five groups by gender and age is shown in Table
II.
1 able il I fequ-jH.v ^s^t *bis!N ti "1 ft, tbikinHi trmn rut il acs* o"
hlM ii3M pii^vMttc il'Hi-, 'KtotvUtw to sfUKki asrt\ acd HtMHfk
'4ru«p
tlu«l.l Mil,- I'tmA \\t\t lemik
gfoy|> it i^'i ' h 1 N ' /• v.llyv r. < 'i °J rV iMaiue J s.'tstl
A 2<> U i> <«« !H 1, 1 IHHl 1 •)".*. Ho
it: . 11. ^ i n» rt- 1
< is »(> 0 s-i j '1 ^N ,1 ml' j In
>•>! f (is i' nil ( li n
D JU lb OMH (•> M u JJ1) HJ
(11 l> 1 1) I fi« 21 t H.i>
!, 12 lit t>«'i 1 ll> >) nil Vi
>M \ 5 H Sp 1 41 2; o'^ ^
'ii«ii iin 121 tt -,0't 1^1 hi" o •)',•: rr.
Drinking water fluoride concentrations were measured in 500 mL samples taken from the 28
wells (1-2 per village) that were the primary source of drinking water. Samples were collected
in pre-cleaned polyethylene bottles and analyzed for fluoride in 2002. The water fluoride
concentrations ranged from 0.3-7.6 ppm, as shown in Table I.
Quantitative exposure data from other sources were not provided. The communities had no
dental fluoride supply or supplement program. According to local custom, eating fluoride-rich
seafood or beverages was rare during childhood, and (fluoride-containing) tea was rarely drunk
fcfr.™ ti,Q o™, ~f * r
-------
ANALYTICAL
METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL
METHODS:
RESULTS:
Dental fluorosis
Chinese populations. Infants are generally breast-fed <4 months, and are given water sweetened
with sugar as early as one month afterbirth (for weaning or supplementary food).
The well water samples were analyzed for fluoride with a fluoride selective electrode (Model
PF-1, Electric and Optic Accessory Factory, Shanghai, China) by the Shaanxi Institute of
Endemic Disease Control. Data were not provided for any other water quality parameters.
Well water fluoride concentrations were determined in 2002. Children from 13 schools were
selected who were 12-13 years old and lifelong residents of villages where the same water
source had been used for the last 13 years. For data analysis, the children were subdivided into
five fluoride exposure groups (A, B, C, D, E) based on their water fluoride concentration, which
ranged from 0.3-7.6 mg/L (Table I).
The children's teeth were examined in 2002. The exams were conducted outside of their
schools, after the teeth were cleaned and dried with cotton balls, using regular chairs, indirect
natural light, and dental mirrors and explorers, per World Health Organization (WHO)
guidelines. One dentist assessed the buccal surface of each permanent tooth in the mouth and
scored dental fluorosis using a modified Thylstrup-Fejerskov (TF) Index (Fejerskov et al.,
1988). Teeth were excluded from analysis if the buccal surface was covered with calculus or if
<50% of the crown had erupted (all 2n molars were excluded). The median of all scored teeth
for an individual was used as that individual's TF score, and was used to evaluate the prevalence
of dental fluorosis (TF >0) and the mean TF score of each fluorosis exposure group. Dental
caries were scored using the DMFS index per WHO guidelines. Missing teeth were considered
to be due to caries if the child had a history of pain. A re-examination after one month of the
four upper incisors of 24 children for fluorosis and of the first molars for dental caries
determined that intra-examiner reliability was good (Cohen's Kappa value of 0.703 and 0.91 for
fluorosis and dental caries, respectively).
To determine whether storage of the drinking water in clay pots (which may either take up or
leach fluoride) affected the prevalence of fluorosis or dental caries, the children were asked to
fill out questionnaires regarding their means of water storage. Of the 477 children, 362
participated from groups A-D; the children in group E were not allowed to participate by their
school teacher.
Dental fluorosis was assessed using a modified Thylstrup-Fejerskov (TF) index, and dental
caries were evaluated using the DMFS index (see Section 2 for description of indices).
The SPSS (Statistical Package for Social Sciences software, version 1 1 .5) PC program was used
for statistical analysis. The gender and age distribution of the subjects were evaluated by the
binomial test. The Chi-square test was used to evaluate the prevalence of dental fluorosis and
dental caries in the five exposure groups. The mean TF score and DMFT were compared with
the univariate general linear model one-way analysis of variance (GLM-ANOVA) and the post-
hoc Bonferroni test. Binary logistic regression analyses were used to estimate the risk of dental
fluorosis (TF score >0) and severe dental fluorosis (TF score >3), with the individual median TF
score as the dependent variable. Significance was defined as P < 0.05.
There was a direct correlation between drinking water fluoride levels and the prevalence of
dental fluorosis (Figures 1 and 2; Table III). The percent of children with dental fluorosis (TF
score >0) increased from 14% in group A to -98% in group E, the percent with a median TF
score >3 increased from ~2% in group A to -95% in group E, and the percent with a TF score
>4 increased from 0% in group A to -60% in group E. Analogously, the mean (of the
individual median) TF scores increased from 0.30 for group A to 4.78 for group E, and were
statistically different between all groups except between groups C and D. Gender had no effect
on the TF scores, whereas the mean TF scores were greater in children of age 13 than 12, the
difference being statistically significant (pO.Ol) for groups B and D (Figure 3).
218
January, 2008
-------
e°-
60 '
Fluaicte group
I i^trrt "t I'ru ^iuKi? "i »,VrH.il tiuoi *is ! If * *T% lt? II *>
* indd 3 lf«
lower
i ^..i im « it. mv t\ii u t mtt i.sn t>f \u.
OD
JH nfOUp
Dental caries (DMFT)
The water fluoride concentration was inversely related to the fraction of children with dental
caries (DMFT > 1), which ranged from 11. 1% in group A to 2.6% in group E (Table IV and
Figure 1). The mean group DMFT values were also inversely related to the water fluoride
concentration, and differed statistically from one another. In fact, children exposed to the
highest fluoride concentration (group E), who had the highest mean TF score (4.78), had the
lowest prevalence of caries among all the groups. The majority (>80%) of the children had only
1 or 2 caries, and 73 of the 83 decayed teeth were first molars.
HlbSc IV. I rt,Tsikffi.c ,*n
JcnMt vines ifi ifec t
Cl ui.t, ,*i"LoK!ms? *u
4iy '! I t
Mcjn-
DMI-.I 1 O»T- I.ppur
lip
Hi
III
l»
ill
Toi.il i, i'> Ul (>.1S 0.1'
* Adjusted for age and feider.
Effect of clay pot water
storage on dental
fluorosis
More than half of the 362 children (from groups A-D) who participated in the survey revealed
that their drinking water was stored in clay pots prior to use. Statistical analysis found no
relationship between the type of container used to store water and the prevalence of dental
fluorosis, although the severity of fluorosis was greater if the water was stored in clay pots
(Tables V and VI). Children from families that drank water stored in clay pots had lower mean
DMFT scores than those who did not, but the difference was not statistically significant.
219
January, 2008
-------
I *h!L V lfre\ ijem-e ul ck'sst«? t!u\>r -MI, U smte " ^ ind 1^,1-
i-b^rcH HsU" Ja-fi'fOi (b^Unklrcn fri-m rui il isv.i4. >IM;,i in1* iprnvsnui.
Tahiti \t \JuiHtL) I mt-sn It ^tm. uuf'+^S t I m pcrrTUn^nt tttittx
m s''is. u} iMwn inmt runl *M*e«i*> ut Mn tru; pnsanve, ( luu, b^
j!u>Tuk £^\'Up .nit11 v*,siiiir'sti^rwu
l:Uoru!e Sn-t,i^ Lit \ltan TF
sa-^uf v.,u^r ?•* SLOR, ^'s'f ll ^-v.-iJuc
t Lnvp^ts
No
5 2 r*0 it OOn
o 2 'h"1
S"1 bi- 0 "(H
n J-t
4 4 ->0 U 0|c<
STUDY AUTHORS'
CONCLUSIONS:
Ruan et al. (2005a) concluded that there was a direct correlation between drinking water fluoride
concentration and the prevalence and severity of dental fluorosis in permanent teeth. Fluorosis
was low grade and uncommon at water fluoride concentrations <0.5 mg/L, whereas >60% of
children drinking water with 0.8-1.4 mg F/L had fluorosis. The authors speculated that the
greater degree of fluorosis in China than in Japan and Western countries at a given water
fluoride concentrations could be in part due to calcium deficiency in the Chinese population. It
was also noted that the DMFT of the Chinese 12-year olds was much lower than that of Western
countries. Gender had no effect on fluorosis. Mean TF scores were greater in children age 13
than 12, although the difference was statistically significant for only two of the five groups (B
andD).
An increase in water fluoride concentration was associated with a statistically significant
reduction in caries. However, the study authors concluded that the harmful effects of fluoride
(not specified; presumably fluorosis) outweigh the benefits of fluoride in their particular
circumstances.
The increase of severity of dental fluorosis for children who drank water stored in clay pots was
of uncertain significance. There are differences in composition of various clays, how clay pots
are made, and in their ability to absorb fluoride from the water (some may leach fluoride into the
water). During the present study, clay pots used for domestic water storage were not specifically
examined for their fluoride interaction characteristics.
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE
NOT FOUND IN NRC
(2006)
Ruan, J.P., Z.L. Wang, Z.Q. Yang et al. 2005b. Dental fluorosis in primary teeth: a study in
rural schoolchildren in Shaanxi Province, China. Int. J. Fed. Dent. 15: 412419.
Fejerskov, O., F. Manjii, A. Baelum and I.J. Moler. 1988. Dental fluorosis; a handbook for
health workers. Copenhagen: Munksgaard (Denmark).
PROFILER'S
REMARKS
SM/1-
16-2007
This was a well-conducted study that clearly showed that the prevalence and severity of
fluorosis increased with the drinking water fluoride concentration. The study also showed that
the prevalence of caries decreased as water fluoride concentration increased. The study authors
downplay the data indicating that the prevalence of caries was lower at the highest fluoride
concentration (group E), which had the highest mean TF score (4.78), than in all other groups.
Gender had no effect on fluorosis, but the effect of age was equivocal, as the differences
between 12 and 13 -year olds the age groups was small. The study authors conclude that the
presence of fluorosis was the children's primary health problem, and that the fluorosis
outweighed the benefits of caries reduction, but do not really explain why they consider this to
be the case.
The same authors concurrently studied fluorosis in primary teeth of 7-8 year-olds in the same
area (Ruan et al. 2005b; see profile). This study also found that fluorosis severity and
prevalence increased with water fluoride concentration, but the severity was slightly greater in
7-year olds than in 8-year olds, irrespective of gender.
The major study drawback is the lack of quantitation of fluoride intake from drinking water and
220
January, 2008
-------
PROFILER'S ESTIM.
NOEL/NOAELfor
fluorosis
PROFILER'S ESTIM.
LOEL/LOAELfor
fluorosis
POTENTIAL
SUITABILITY FOR
DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
other possible sources. The study authors noted possible differences in susceptibility to
fluorosis in the Chinese population due to dietary factors such as calcium deficiency. Other
drawbacks include the use of natural light to examine the children, and lack of data regarding
how many teeth/child were examined, and lack of quantitation of fluorosis and dental caries on a
per tooth basis, and by tooth type. It would have been helpful for subsequent analysis if these
data had all been presented as numeric values instead of bar graphs.
The NOAEL for fluorosis was 0.3-0.6 mg/L (mean=0.4 mg/L) in the drinking water (exposure
group A), which caused no fluorosis in 86% of the children and only mild fluorosis (TF <4, no
pitting) in the remaining 14%. A concentration at which no fluorosis occurred (NOEL) was not
identified.
The LOAEL for fluorosis was 0.8-1.4 mg/L (mean=1.0 mg/L), corresponding to exposure group
B, for which 40% had TF=0; -58% had TF=l-4, and -2% had TF>4. This group was chosen as
the LOAEL because some subjects had severe dental fluorosis (pitting occurred).
Not suitable ( ), Poor (), Medium (x), Strong ()
The study shows a direct correlation between water fluoride concentration and the incidence of
fluorosis or dental caries. The major study drawback is the lack of quantitation of fluoride
intake from drinking water and other possible sources, but estimates can be made of the
children's fluoride intake based on age and expected water consumption and dietary profile.
The data are also amenable to analysis of the dose-response relationship between water fluoride
concentration and the prevalence of dental caries.
Dental fluorosis; dental caries
221
January, 2008
-------
Ruan, J.P., Z.L. Wang, Z.Q. Yang, Bardsen, A., Astrem, A., and Bjorvatn, K. 2005b. Dental
fluorosis in primary teeth: a study in rural schoolchildren in Shaanxi Province, China. Int.
J. Fed. Dent. 15: 412-419.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
Dental fluorosis in primary teeth
Case control, retrospective
China/ Shaanxi Province: 472 schoolchildren from two rural areas (Baoji County and
Jingbian County). The province is in northwestern China between 3 1-39°N and 105-
1 10°E. The children were age 7-8, approximately half male and half female, and attended
13 village primary schools. The communities had comparable socioeconomic standards.
The drinking water source was mainly groundwater obtained from 20-30 m deep wells.
The source of village water was unchanged over the last 9 years. No other source of
fluoride was known in these communities. Shaanxi Province is known for the presence of
endemic dental and skeletal fluorosis (Liu et al 1999).
There was no control population per se, rather, children were subdivided into four fluoride
exposure groups, which were all compared to one another.
Lifetime, from prior to birth through the study exam at ages 7-8.
Schoolchildren aged 7-8 from two rural areas in Shaanxi Province, China, who obtained
their lifetime water supply from one of 27 wells. The children were subdivided into four
fluoride exposure groups (A, B, C, D) based on their well water fluoride concentration, as
shown in Table 1. The distribution of the children in the four groups by sex and age is
shown in Table 2.
Table 1. Fluoride tutKenliation in litmlcing »atn
rioorule euiKHitratien imgl-'i
frTcrap \Vlh IB! Minima m Maximum Mean
A 12 0-3 Hi i>*
I 5 1-2 2-0 1 5
C 9 24 3-1 3 2
D I 74 7-6 7h
Total 27 0-3 74 2 (I
fabls 2, Frequency distnbutKso ot participant^ bj «tufer and sse
7-Kir-ddr 8-,cir-dds
Oroup Mile \i l*i] kralt |ni*>] P-vitar Male [n l"rl] kmaJc [* i'f-!] f-fslic letii
\ MM^Si Siuihlj 04.X Mimi -K^fiOi 141 2M
B H 140-5) 22f515! S45i 1-00 19
tail 130 131 1-00 KB 104 0-89 472
Drinking water fluoride concentrations were obtained for 500 mL samples taken from the
27 wells (1-2 per village) that were the primary source of drinking water for the 472
children. Samples were collected and analyzed for fluoride during the summer and fall of
2002. The water fluoride concentrations ranged from 0.3-7.6 ppm.
One of the criteria for selection of the village school for study was that "apart from the
drinking water, no fluoride source should be available." No further information was
provided regarding other possible sources of fluoride exposure.
222
January, 2008
-------
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
The well water samples were analyzed for fluoride with a fluoride selective electrode
(Model PF-1, Electric and Optic Accessory Factory, Shanghai, China) by the Shaanxi
Institute of Endemic Disease Control. Data were not provided for any other water quality
parameters.
Well water fluoride concentrations were determined in the summer and fall of 2002, and
used as a basis for selecting the participating schools. Children were selected from the
schools who were 7-8 years old and lifelong residents of villages where the same water
source had been used for the last 9 years. For data analysis, the children were subdivided
into four fluoride exposure groups (A, B, C, D) based on their water fluoride concentration,
which ranged from 0.3-7.6 mg/L (Table 1).
The children's teeth were examined in 2002. The exams were conducted outside of their
schools, after the teeth were cleaned and dried with cotton balls, using regular chairs,
indirect natural light, and dental mirrors and explorers, per World Health Organization
guidelines. One dentist assessed the buccal surface of each primary tooth in the mouth and
scored dental fluorosis using a modified Thylstrup-Fejerskov Index (TFI). Intra-examiner
reliability was moderate (Cohen's kappa of 0.58), as determined by a repeat examination of
29 children after one month.
The TFI score of the upper right second molar was used as an individual's overall TFI
score. The mandible and maxilla TFI scores were those of the right upper and lower
second molars. The dentition distribution of the TFI scores was based on the scores for all
of the teeth of a given individual.
Dental fluorosis was assessed using a modified Thylstrup-Fejerskov Index (TFI) (see
Section 2 for description).
The SPSS (Statistical Package for Social Sciences software version 11.5) computer
program was used for statistical analysis. The gender distribution of the subjects was
evaluated by the Binomial test. The Chi-square test was used to evaluate the prevalence of
dental fluorosis (TFI score >1) in the four exposure groups. The KruskalWallis /f-test and
the Mann- Whitney U-test were used to evaluate the median of the TFI scores in the four
groups by gender and age. The percentage of dental fluorosis of the maxillary and
mandibular dentitions was compared by the Wilcoxon signed-rank test. Significance was
defined as P< 0.05.
As shown in Table 3, there was a direct correlation between dental fluorosis and the
fluoride level in the drinking water. The percent of children with dental fluorosis (TFI
score >1) increased from 6.2% in group A (0.3-1.0 mg/L fluoride) to 96.6% in group D (7.6
mg/L fluoride). Similarly, the median and 25th and 75th percentile TFI scores increased
from zero for group A to 4 for group D. The median TFI scores were statistically different
between all groups except between groups B and C.
Tafek J The prevalence and seventy ff derrtsJ fluerosis in primary teethe Tbyl^trup-FefsrFskn^ Inde^ iTFIi
Twenivtrtth Se\?nty-flfih
Gruup Natnbci Medbon pcjx'tiitile perwat&te Niftier Pscc-siige
A 2S 0« g 0 14 6-1
1 S3 «*' fi I 1& 31 J
C 135 II" 11 4 S4 4M
D 2J 4« 4 4 a 964
lettl 471 il U 1 122 2S§
^The jyn&tkantl]! dilferen? meditui ui !f--*eme*i w« tound bei^coi ail ymnp
*wNo •tatnticifK RigFoiseaul different medlao ot TT-^c^res ¥ §£ fnynd bft^ecn gjcup B sod C
There were no statistically significant effects of gender on the median and 25th and 75th
percentile TFI scores of the four groups (Table 4). Age had a significant effect on the
223
January, 2008
-------
TFI score only for fluoride group C (2.1-3.8 mg/L), as the 7-year olds had higher 75th
percentile scores than the 8-year olds.
Ts*bte4 Mcdun of Th^btri!p-l'ejen4cw lnds\ aTFIi woie^ lar aft sod gender Fus tfe IW fluonde grou^
tTOiyp A Ijfuup B Gimp € Grisyp D
VanoHc i M«hen
Grader,
mils 120 9
femrfe 109 0
Age (yean}:
1 129 0
« ?§ §
P^rcesi-tit? PesxTnfele IVrcenhle f^Kcoi]!?
.25. m x Median ^5, "5) , Median l2\ 75| , Mediin l;25, TO
0, 1 38 0 0, 3 6§ 0 0, 4 13 4 4, 4
0, 1 « 0 I, 3 69 0 0, 4 IS 4 4, 4
0, 1 37 0 0, 3 77 0» 0, 4 18 4 4, 4-Z!
0, i 46 0 0,1 Si 0 0, 2-23 11 4 4,5
*Pl»trap-tfjcr*Lov Index (TFIi »ctsre and e«?up
TFI Kort
Groip 0
A 934
i 68-7
C 60-0
D 34
1-2 V-4 5-ti >?
4-0 :•: ltd oo
34 :i 7 nn ia
i-7 2M 44 M
0-0 724 207 34
The percentage of dental fluorosis was slightly greater in the upper jaw (maxillary teeth)
than in the lower jaw (mandibular teeth) for all four fluoride exposure groups, although
the difference was statistically significant for only group A. Fluorosis was distributed
symmetrically in both the mandible and the maxilla in all groups. The second primary
molar was the most fluorosed tooth, then the first molar, then the canines, the central
incisors, and lastly the lateral incisors. The results are depicted in Figures 1 and 2.
PROFILER'S COMMENT: The tooth numbering system utilized in Figs 1 and 2 is that
of the Federation Dentaire Internationale (FDI) World Dental Federation ISO-3950,
where Is are central incisors, 2s are laterals, etc. The FDI "baby teeth" notation employs
51-55 for upper R, 61-65 for upper L, 81-85 for lower Rand 7 1-75 for lower L.
100 f
80 [• "J I- " - :ii ;
a. 1 -'),!> <-_ft !
f 60 h
®
S 40
a. . 4_,
20
Fsg. 1. rronipsmon vt the p*
ypjsr rad loavi )3¥ b> iluoi
4
•v1* .
-.^ j '
-r™w^ — ^ I ^ i j
i r^f f |v. i
i C _i
'"in. l|
romi^s if dgnlrf §Marasis to Ife
ide ^ «!(•
224 January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date
REMARKS SM/1-12-
2007
PROFILER'S ESTIM.
NOEL/NOAEL for fluorosis
90-
f 99-
is
0-
0-
-JO-
&
0.
-100-
^U 'Up
i>6 b* 51 41 M «1 & S3 84 6S
T*«lh
81 84 83 &> 81 71 7r "3 7* 7S
'i^esfaK? oi dental flu.irw^ act 'Hdii.g to t
-»- (5 «If> *
G*0 !f> B
-*- (Vis* A
tlfn ip B
Ruan et al. (2005b) concluded that there was a positive relationship between the prevalence
and severity of dental fluorosis in primary teeth and drinking water fluoride concentration.
Fluorosis was distributed symmetrically throughout the mouth, with the second molar being
the most affected, and the lateral incisors being the least affected. Gender had no effect on
fluorosis, and age had a small equivocal effect. Because most of the TFI scores were 3-4,
the authors speculated that there may be a dental fluorosis cut-off point above which the
ameloblasts lose their ability to produce adequate enamel.
Ruan et al. (2005b) asserted that fluorosis in primary teeth is an early warning of excessive
fluoride exposure, and provides a basis for intervention to prevent fluorosis of the
permanent teeth. The authors encourage feeding infants breast milk because it is low in
fluoride, irrespective of the mother's fluoride intake.
Ruan, J.P., Z.Q. Yang, Z.L. Wang, et al. 2005a. Dental fluorosis and dental caries in
permanent teeth: rural schoolchildren in high-fluoride areas in the Shaanxi province, China.
ActaOdont. Scand. 63: 258-265.
Liu, X.L. 1999. A brief introduction to local disease control in Shaanxi Province.
Endemic Diseases Bulletin 14: 72-73.
This was a well-conducted study that clearly showed that the prevalence and severity of
fluorosis increased with the drinking water fluoride concentration. The lack of a
statistically significant difference in the TFI scores of groups B and C may be due to the
fact that these groups were exposed to very similar water fluoride concentrations. The
study also showed that the distribution of fluorosis was symmetric in the mouth, that the
second molars were the most susceptible to fluorosis, and that gender had no effect on
fluorosis. A conclusion regarding the effect of age could not be made, not surprisingly,
since the children were close in age.
The same authors concurrently studied fluorosis in permanent teeth of 12-13 year-olds in
the same area (Ruan et al. 2005a; see profile). This study also found that fluorosis severity
and prevalence increased with water fluoride concentration, but the severity was slightly
greater in 13 -year olds than in 12-year olds, irrespective of gender.
Drawbacks of the study include (1) the less-than-optimal intra-examiner reliability
(Cohen's kappa of 0.58), as determined by a repeat examination of 29 children after one
month, (2) the use of natural light to examine the children, and (3) the lack of quantitation
of the children's fluoride intake, including that from other sources besides drinking water.
The NOAEL for fluorosis was 0.3 to <1 .0 mg/L in the drinking water, corresponding to
exposure group A (0.3-1.0 mg/L, mean=0.6 mg/L). At <1.0 ppm, fluorosis severity was
225
January, 2008
-------
PROFILER'S ESTIM.
LOEL/ LOAEL for fluorosis
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
limited to mild or moderate (score 1-4), and 93.6% of the subjects had a TFI score of 0. A
concentration at which no fluorosis occurred (NOEL) was not identified.
The LOAEL for fluorosis was 1.2-2.0 mg/L (mean=1.5 mg/L), corresponding to exposure
group B, for which TFI=0 of 68.7%, TFI=l-4 of 30.1%, and TFI >7 of 1.2% of the
subjects. This group was chosen as the LOAEL because some subjects had severe dental
fluorosis.
Not suitable ( ), Poor (), Medium (x), Strong ()
The study shows a clear dose-response for water fluoride concentration vs. dental fluorosis,
despite some drawbacks. These include moderate intra-examiner reliability, the use of
natural light for examination, and lack of quantitation of fluoride intake from drinking
water and other possible sources. Estimates can be made of the children's fluoride intake
based on age and expected water consumption and dietary profile.
Dental fluorosis in primary teeth
226
January, 2008
-------
Rwenyonyi, C.M, K. Bjorvatn, J.M. Birkeland and O. Haugejorden. 1999. Altitude as a risk
indicator of dental fluorosis in children residing in areas with 0.5 and 2.5 mg fluoride per litre in
drinking water. Caries Res 33:267-274.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT
ANALYTICAL
METHODS:
Dental fluorosis (permanent teeth)
Cross sectional survey
Uganda/Rift Valley region: Children aged 10-14 years (mean 12.2 years) from three rural
schools of Kisoro district and two schools in Kasese district, both mountainous areas of
Uganda. Within the Kisoro district, the Mutolere/Kagera (n=163) and Kabindi (n=155) areas
were at altitudes of 1,750 and 2,800 m, respectively. Within the Kasese district, the Mpondwe
(n=81) and Kyabayenze (n=82) areas were at altitudes of 900 and 2,200 m, respectively.
Children had to satisfy the following criteria to participate in the study: 1) be born between
1982 and 1987 (aged 10-14 years) and raised in the village where they presently lived; 2)
should have not been absent from the village for more than one month in any calendar year
and should have used drinking water from the same source during the first six years of life.
None.
Birth to 10-14 years. Study was conducted in 1996 and 1997.
Within the Kisoro district, the water fluoride content was 2.5 mg/L (range 2.41-2.60; n=17).
Within the Kasese district, the water fluoride content was 0.5 mg/L (range 0.47-0.5 1; n=6).
The three areas in the Kisoro district were supplied with piped water from the same spring.
The two areas in Kasese district obtained water from the same river either directly or through
a piped system. Samples of drinking water in the Kisoro and Kasese areas were analyzed for
fluoride. Questionnaires were used to determine fluoride from other sources, including the
following: boiling of drinking water, clay pots (for water storage), trona (calcium carbonate
used as food additive), tea, infant formula, vegetarian diet, milk and fluoride toothpaste. The
fluoride exposure from liquid (PEL) for each child was calculated from the reported number
of cups of liquid consumed per day and the fluoride levels in the drinking water.
Table 1 was copied directly from Rwenyonyi et al. (1999).
T»biB 1. Altitude of residence. liquid tnulf
(rrean * SD»>. and FLi (mean ± Sf>! tiy dis-
trict
rxuir KaMTMr (0 !S nig F/h Kiwrof2.5 wg F/IJ
^IjwJwe Kyat»cn*'t: Miiiolfir/Xafcra K^triitji
in- HI) (n = 82t f-iw 163) (niJSSJ
AiirsjJt'.ir "«» ?,2fX> I 7VS MWO
I H;UI,! '"«.;U fie- '< ?±-H 1 »-' h 1 *-i'> i 5.fti.5
J-i L n-y. fVd«y Dh+03 II6+U3 1 N± ! 2 .V?*M
Analyses of the drinking water were conducted using a fluoride ion combination electrode.
227
January, 2008
-------
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL
METHODS:
RESULTS:
The objective of the study was to assess the association between altitude and dental fluorosis
among Ugandan children in two fluoride districts while controlling for other factors related to
fluorosis. All children who satisfied the criteria were lined up in the school yard, with two
lines according to gender. Every third child was selected for the study, totalling 491 children.
Ten children were excluded, 9 because of non-continuous residence in the villages and 1 who
could not be traced for an interview. All the children were ethnic Bantu Africans and their
socio-economic backgrounds appeared to be the same. There was no significant difference in
the distribution of children according to gender. Written consent to participate in the study was
given by the children's parents.
A random sample of 481 children aged 10-14 years was examined for fluorosis using the
Thylstrup and Fejerskov index (TF; see NRC, 2006, pages 88-89). A few days after the
clinical examination, the child and his/her mother were interviewed according to a structured
questionnaire to determine other sources of fluoride, including use of the following: boiling of
drinking water, clay pots (for water storage), trona (calcium carbonate used as food additive),
tea, infant formula, vegetarian diet, milk and fluoride toothpaste.
The prevalence of dental fluorosis was determined by clinical examination. One trained
dentist examined the children under field conditions. The child was seated outside the school
building and only sunlight was used for illumination. The surfaces of the permanent teeth
were cleaned and dried with cotton balls prior to the examination. For all permanent teeth with
at least 50% of the crown erupted, the severity of the fluorosis was assessed on the
buccal/labial surfaces using the modified TF index. About 10% of the children (n=46) had
their upper right central incisor re-examined a day later for a reliability test. There was no
evidence of a systematic error. A few days after the clinical examination, the child and his/her
mother were interviewed according to a structured questionnaire to determine other sources of
fluoride, including boiling water, clay pots (for water storage), trona (calcium carbonate used
as food additive), tea, infant formula, vegetarian diet, milk and fluoride toothpaste.
Chi-square statistics were used when comparing frequency distribution of children on the
basis of the prevalence and severity of dental fluorosis, age, gender and other variables (see
Table 2). Student's t test for paired observations was used to check for systematic errors in TF
scores. Spearman's rank correlation coefficient (r) was used to study the bivariate association
between variables. Student's t test for independent samples was used to test whether r was
significantly different from zero and to test differences between means of quantitative
variables. Stepwise multiple linear regression analyses were used to control for confounding
and to identify factors explaining variation in the percentage of teeth affected by fluorosis.
Multiple logistic regression analyses were used to estimate the magnitude of risk of
developing dental fluorosis.
The distribution of children according to independent variables is presented in Table 2 taken
directly from Rwenyonyi, 1999.
228
January, 2008
-------
* ' '*'• uiOirCii -*!> ' "K»ttHR«-8J) ],Mlln,(n« (Ml .'K(M m (n
l!,,!i-c,>t ,". (•<"" '<• l-i '*" l(l '.i(te<
.'I'JB| .'Mil HI (Ml 74 [4M
-ni7ii ^c '*ii(."f »" 4-'i ^s si -.it
h (MI) I'll li) MS)
KS HiW) tW.S'ii S IWl'rtj
jutwik H> IfiS'l ,'»li'l IMI'»:
n,, M. -tv'i v. , 'IN NS
Study results in Tables 3 through 7 and Figures 1-3 are shown directly from Rwenyonyi,
1999.
Table 3, 111 ''4 »wt. ± Sl»i «»i IF icuic '- i. utmnlirt «' Oi L"a-n.
iiiiiiuik* of rflsidi'ntt 3«J 1 coii.*• Meat 10.1 in ih; dr«i'kin^ water h-« J!"
Ht4l*w! ir-g F/i Akittufc.ff rhiUrcr, u TPf, "
(nwan i SD)
WMI 8!
8?
Kivoro 2.5 I,"HO 163
155
AH 4Ki 554_»44(f',
229 January, 2008
-------
Tabl« 4. ,*»pr;ntnan K mud tfitiddinin vocl'iivieui (r.K btlwwrt tee,
a-!-.1)?' fl.L. liftuuik* »it»U s',*,»f;iff of limiting wawr in earthenware
t™f.K ami Til uf IC,»MM' J sirii'i with 0 5 tut- H! ir dunlin;? wito
"Iff- Aff (Jemk*r I hL Aliiludc
Age ' « 05
r Oil 0 SK
fr - Hi I) (H •- 1
FEf 0^1* -00>
fn » t4»j ts-r - i-
002
f,i,i> (103 -Oil G.ftf
in- l(Sl> <»i = l*»1> un-14.1) 'PS
•>,-,< OH.*, |1|t«'OOK" «1 17 -0 It -ftOg
»*>« J.w.i;!.- (life- -tMW -till.' 0^2' 00' O.W 0.05
-p<0lrt, "p«- litll, >(K.ii(iiH (1 i-.II
230 January, 2008
-------
Tabla ft, Ste;M\isc multiple linear rejwssHVi jinalj .-.cs sh*wn
R2^,,r a^J (*•' aJ(uveJ li*i I'l.L. M<»MI'<: ul dunking water in e;itrlwii-
«,are funs, alt.luck' in1";*-!! formula «.e ot I K-othftav*.1 .inJ ves.'.-tafi-
en TPI; bv t
independent
variable
1-hL
Alium
\Va't»
Vcg«'i3
Jftf.int fnrnnilt
(n= 143,
!) 5'iif F/0
D.276
o U4»;
0 'i.w
(n * 3<
OM?
R1 ad rustc d
0«'6
0 5
All
« 36^
U.UI'J
T«bJ» 7, Ldttr.Mic rejsrONiun jna'ysoi shnwjnjt OR wml 95*^ cit;ilV
ik-r.cf interval (CI) Iw jtiniwl tlui>n>sts (I! - "I F *. 3, t = If I* 1| ;t%s<>-
cuied with iri)teptfnik"ti variahiev: afi-iadt «f «CMI|^IKC lU = tow,
1 « high: WOO v*. 2.2CXI m in Kasrse and 1.750 vs. 3,8(X) itt in Kiso-
ro>; I-EL, (0 «• in KIHIHI), worage rmwl.t
(CS ss fid, ] - yf*,) hy df\tf 1U
Independent
= 1431
OR C95*' CI}
Ff-!L
Aloiudc
65(24
2,9 113
M}.6'I5 I -487)
5 ! (16-1')9)
231
January, 2008
-------
-o
300 m
Tf-4
KKH) m
r~i
f
Fig, 1. Uisinhulton ol children occnrdnt^ ID ihc hlj|hc>«l IF scores
recorded, h coneeritaiifw i»1 Ow dnnkir>p wat« ;imj alsiiwde of fcs>-
dcncc by (ftsSnu,
P«t«i
Fig, 2. Cluwuurtnx' tictjuency disitribmion c( children according t
pruj'urfion of ie«I'- per c^iiid exhibiting demjl tluiMv^m «f TF icorts
>'*, »i dit'lcrcol afiifudcs of reMiieiifc unc( F conccnuiitioiis of Ac
232
January, 2008
-------
•"2800m,
-•-•1750m.
•*~ 2KB m; 0 5 fig FH
BGOm, 0.5 me f A
Fig, 3. t!unwl,(ii\c fi'1'tjuefH.y
fwpiMww of (coll) p« rhnJ c
•^S, n{ difU'teru aliitutfc*, tit
({finking witter
listnt>uii»n t»l diiiiiit'n .im>r Jin^ tn the
thtbtiinp ilcwu! Ounrotis t>| Tl srarfs
Hkwnci- ami I-' a>m.cr,!fai»»m of iftc
STUDY AUTHORS'
CONCLUSIONS:
The study authors concluded that most of the variance in the prevalence and severity of dental
fluorosis was explained by the fluoride intake from liquid, but altitude remained a significant
risk indicator after controlling for the effect of other potential confounding factors by multiple
and logistic regression analyses.
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
None
PROFILER'S
REMARKS
Initials/Date
VAD/03-07-
07
The study results are not representative of the U.S. population since the study was conducted
in Uganda. The number of participating children in the Kisoro district was almost twice the
number in the Kasese district which could have biased the study. No reliability tests of the
interviews were conducted because of the field conditions and therefore, the possibility of a
recall bias cannot be ruled out. The examiner was not blinded to the children's district and
hence fluoride exposure level. The study report does not indicate when the drinking water
analyses were done; seasonal variations are known to occur.
The technical reviewer agrees with the profiler's estimated LOAEL but the applicability to the
U.S. population is limited due to some outside sources of fluoride not observed in the U.S.
(i.e. use of trona, storing water in clay pots) and the high altitude of the areas profiled.
PROFILER'S ESTEM.
NOAEL
The study design did not identify a no-fluorosis intake dose.
PROFILER'S ESTEM.
LOAEL
In this study, the LOAEL was 0.5 mg/L. In the district with 0.5 mg/L fluoride in the drinking
water, 25% of children at 900 m had dental fluorosis (TF score >1 on at least one tooth),
whereas 45% of children living at 2,200 m were affected.
SUITABILITY FOR DOSE
RESPONSE MODELING
Not suitable,O; Poor Q; Medium (X); Strong (_)
233
January, 2008
-------
CRITICAL EFFECTS:
Although the study does provide dose response data,
therefore, not suitable for modeling.
it is presented in graphical form and
Dental fluorosis (permanent teeth)
234
January, 2008
-------
Selwitz, R.H., R.E. Nowjack-Raymer, A. Kingman, and W.S. Driscoll. 1998. Dental caries and
dental fluorosis among school children who were lifelong residents of communities
having either low or optimal levels of fluoride in drinking water. J. Public Health Dent.
58(l):28-35.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
PARAMETERS
MONITORED:
STATISTICAL METHODS:
Dental fluorosis; dental caries
Cross-sectional survey of dental caries and dental fluorosis; follow-up of similar studies
conducted in 1980; 1985 and 1990 (Horowitz et al 1984; Driscoll et al 1983, 1986;
Heifetz et al 1988; Selwitz et al 1995).
US/Illinois; children in two age groups, 8-10 yr (86 males and 81 females) and 13-16 yr
(45 males and 49 females) residing in Kewanee.
US/Nebraska: children in two age groups, 8-10 yr and 13-16 yr, residing in Holdrege,
and Broken Bow, NE. Like the Illinois communities, the two Nebraska communities
were small, rural Midwestern towns. Each had a per capita income of approximately
$15,000, an agricultural economy, the same number of local dental practitioners, and
similar percentages of high school graduates entering college.
Lifetime (8-10 yr; 13-16 yr), up until the dental examinations which were conducted in
the Spring (NE) or Fall (Kewanee, IL) of 1990.
Drinking water fluoride concentration was 1 ppm in Kewanee, IL, and <0.3 ppm in
Holdrege and Broken Bow, NE.
Information was obtained from parents by questionnaire concerning the use of fluoride
toothpaste, prescription fluoride drops or tablets, and professional fluoride dental
treatments.
Based on information provided in an earlier report (Selwitz et al 1995), mean fluoride
water concentrations were determined by averaging all available readings. The optimal
water fluoride level for Kewanee, IL was reported to be 1 ppm. In Nebraska, 76 of the
children received their drinking water from private wells; random water samples were
analyzed with a fluoride-sensitive electrode (Orion Research, Inc.) for 62% of the wells
to verify that the fluoride levels were negligible.
Dental caries was assessed using the DMFS scoring system; dental fluorosis was
evaluated with TSIF scoring system (see Section 2 for descriptions of scoring systems).
Differences in mean DMFS scores of participants by community were tested for
statistical significance using the least square means option under the SAS (Statistical
Analysis System) general linear models procedure (SAS 1990). The chi-square test of
homogeneity was used to compare differences in the prevalence of dental sealants
among the communities and differences in responses to questions regarding the
participants' fluoride histories (SAS 1990). For dental fluorosis, the primary subject-
based summary measure used in the statistical analyses consisted of the percentage of
fluorosed surfaces per subject. Mean scores for this variable (MPFS) were computed
for subjects in the three communities. Fixed effects ANOVA models were used to
make comparisons among the subgroups. The LSMEANS procedure in SAS, which
adjusts the group means for confounders present in the model, was used to compare
adjusted means for statistical significance (SAS 1990). For individual comparisons, an
a = 0.02 value was used to control the overall experiment-wise Type I error rate as a
compromise between the more stringent Bonferroni o/k value and the unadjusted a =
0.05 value. This procedure adjusts for multiple comparisons, but retains the property
of better power for conducting individual comparisons (Kleinbaum et al 1988). All
235
January, 2008
-------
RESULTS:
Caries
Dental fluorosis
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date
REMARKS DMO/1/12/07
levels of significance reported are calculated P-values.
The mean DMFS score adjusted for age, sealant presence, and fluoride use was
significantly lower in Kewanee (1.8) than was the adjusted mean caries score in either
Holdrege (2.9) or Broken Bow (3.6) (see Table 2, copied directly from Selwitz et al
1998).
TABLE t
Compiriioni of Metn DMES Scores for All PHtidprato in Kewanee, Broken Bow, and HoUnge, M90
Age-adjusted Multivariable™ % 98% Confidence Interval for Difference in§
Mam No, adjusted Mean No. Increase Multivariable-adjusted Means
. — --- - aom -
Communities n DMFS (SB* DMFS (SRt K£J BB HO BB&HO
Kewaiwe 260 1.9 (.205 1.8 (.23 - (0.93,2.79) CQ.21,2.13) (0,76,2,30}
Holdrege 128 2.6 (39) 2.9 (.15) 61.1 (-0.45,1.83)1
Broken Bow 107 3,7 (.31) 3.6 (.34! 100,0
Broken Bow & 235 3.1 (.221 3.3 (.241 833
Holdrege
*Mean DMFS scares have been age-adjusted; numbers in parentheses axe standard errors of meafi.
tMean DMF5 scares hsv€ been adfsstsd forage, sealant presence, reported use of dietary fluoride supplements, and reported use of professionally
applied topical fltiortde («=4$3)>
JPereerst werease In mulsivariabte-«ijusted mean l?MF$ score; conwiwisti^s: K&Kewwee lopteml wal« fiuondatsorO, HO=Hoidrege(<03ppm
F), BB=Brokin Bow <<0,3 ppm F).
&Tbe first thrw Cl-a)1(X)cci«ridepee mtervalsptewntesl are fw the dif teener in tnuMwinable-adjysted mvm DMFS scores between KEasid BE,. KE
and HO, and between KE md BB t HQ nspeoively.
^(1~«)100 conMence at«3¥al for dkffecTM in mu3l3var3ab!«-adsusted me&n DMFS scores between HO 3Rd BB.
The mean percent of fluorosed tooth surfaces per person, adjusted for age and use of
dietary fluoride supplements, was similar in the three communities (approximately
15%); more than 80% of tooth surfaces in all participants were fluorosis-free.
TABLE 4
Percent Distillation of TSIF Scores for Participants by Age Group and Cmnmtinity, 1990
., . .. , % Distribution ofTSIF Scores „ _ ,
Nn of Nn nf % ^up-fa^^*
Community (Water Fluoride Level) Children Surfaces 0123 1—7 Fluorosed
Participants 8-10 years of age (age group 1}
Kewaneetaptimal water F> 167 4,867 81,4 14.4 2.8 1.3 O.Ot 18.5
Holdrege (<0.3 pprrt F! 104 2,956 81.7 !2.S 3.4 2.3 0,1 18,4
Broken Bow <<0.3 ppm F) 4? 1,424 82.3 15.2 2.2 0.3 0.0 17,7
Participants 13-16 years of age (age group 2}
Kewanec 93 6,203 83.0 13,1 1.6 0.3 OJ 15.1
Holdrege 24 1,447 97,9 1,9 0,2 0.0 0.0 2,1
Broken Bow 60 3,748 90.9 8.1 0,7 0.4 00 9,2
"Percent surfaces fluorosed across aE sybfecte,
t'Two airfares were affected
Findings from the present study suggest that water fluoridation still is beneficial and
that dental sealants can play a significant role in preventing dental caries. In addition,
findings from this survey appear to support the premise that the difference in dental
fluorosis prevalence between fluoridated and nonfluoridated communities has
narrowed considerably in recent years.
SAS Institute Inc. 1990. SAS/STAT user's guide. Version 6.4th ed. Vol2. Gary, NC:
SAS Institute, 1990:891-6.
SAS Institute Inc. 1990. SAS procedures guide. Version 6. 3rd ed. Gary, NC: SAS
Institute, 1990:325-36.
Kleinbaum DG, Kupper LL, Muller KE. Applied regression analysis and other
multivariable methods. 2nd ed. Boston: PWS-Kent Publishing Co.
Although severe fluorosis was documented in a small percentage of the study
populations, insufficient data were provided to correlate severe fluorosis with a
236
January, 2008
-------
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
significant increase in caries.
The range in fluoride concentrations in drinking water was insufficient to correlate
fluoride concentrations in drinking water with fluorosis
Insufficient data
Insufficient data
Not suitable (X), Poor ( ), Medium ( ), Strong ( )
Dental fluorosis and caries.
237
January, 2008
-------
Selwitz, R.H., R.E. Nowjack-Raymer, A. Kingman, and W.S. Driscoll. 1995. Prevalence of dental
caries and dental fluorosis in areas with optimal and above-optimal water fluoride
concentrations: A 10-year follow-up survey. J. Public Health Dent. 55(2):85-93.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT
Dental fluorosis; dental caries
Cross-sectional survey of dental caries and dental fluorosis in children residing in Illinois in
1990; follow-up of similar studies conducted in 1980 and 1985 (Horowitz et al 1984; Driscoll
et al 1983, 1986; Heifetz et al 1988).
US/Illinois; children in two age groups, 8-10 yr (369 children) and 14-16 yr (188 children).
All of the 14-16 yrolds examined in 1990 had been participants in the 1985 study as 8-10 yr
olds.
None
8-10 yrs and 14-16 yrs (examinations conducted in October, 1990).
Seven study sites were grouped into four categories of exposure:
TAB! F i
^o ni TJ \ *tA ***" v ^.uit^-'.rt o- prT \a a* V^eiYuiifv!
Opt n «.s ^lorus-UM ^-A **.< lw 8 *0 Sx t. f*t J^^r >UiJ 4 A
K( ,-i t m i,, ii Ml * it
1 38 3: M7
Nil tru i'2\ civil ) 2.08 1.95 M 48 34 13
F 58 42 16
Ab>injp)an,E!ntwoad(3X optima!) 2,«7 2.70 M 54 33 21
F &3 36 2?
Bushnell, Ijava, and TaWc Grove 3,89 3.59 M 39 29 10
(4X optima)) F 38 28 ",0
Total 3tS 188
The study participants had lived continuously in their communities since birth, and had
always used the community water supply as their primary drinking water source.
PROFILER'S NOTE: Water supply of Bushnell (in 4X category) underwent alteration in
1982 with addition of lime softening process, which resulted in change to fluoride content of
distributed supply (from 3.8 ppm to average of 2.5 ppm observed at time of 1990
examinations). Thus, children ingesting water from the Bushnell system had not been
exposed to 4X water since 1982 (approx. 8 years at the time of the 1990 exam).
Children in the 8-10 year group were exposed to water with a concentration of 2.5 mg/L for
either all of their lives or 8 out of ten years.
The children in the 14-16 year group had been exposed to water with 3 .89 mg/L from birth to
6 years or birth to 8 years. If enamel formation is complete by 8 or 9 years, then they should
be considered to be exposed to 3 .89 ppm for the dose response analysis while the younger
group should be considered to be exposed to 2.5 ppm for the most recent analysis.
Dental exams took place in local schools with use of portable dental chairs, artificial lights
and plane surface mouth mirrors. TSIF was determined by the same dentists who had
performed these assessments in earlier surveys; to determine level of inter-examiner
agreement, a 12% random sample received duplicate exams. Dental caries criteria were those
of the ADA.
No attempt made to quantify alternate sources of fluoride such as mouth rinses, dentifrices,
238
January, 2008
-------
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental Fluorosis
topical applications, and food supply; authors agree that these sources play a major role in an
individual's composite fluoride intake.
Mean fluoride water concentrations were determined by averaging all available readings for
each community obtained from state and local water officials. The optimal water fluoride
level for that geographic area was reported to be 1 ppm.
Dental fluorosis and caries incidence were evaluated in 369 children 8-10 yrs old and 188
children 14-16 yrs old from seven communities in Illinois having different levels of fluoride
in drinking water (mean values ranging from 1.06 ppm to 3. 89 ppm in 1964- 1980 and from
1.01 ppm to 3.59ppmin 1974-1990). Comparisons were made using acceptable methods of
diagnosis (TSIF index for fluorosis, and DMFS index for caries) and statistical analysis (two
way ANOVA). Comparisons were made with data from the same populations obtained in
1980 and 1985.
Tooth surface index of fluorosis (TSIF) was utilized in the examinations (see Section 2
description). Caries incidence was determined by the decayed, missing and filled surface
index (DMFS index). Radiographs were not taken.
Differences in mean DMFS were tested using analysis of covariance, adjusting for age. The
significance level for interpretation of calculated P-values was adjusted for multiple
comparisons using the Bonferroni procedure (Bohannan et al 1984). Mean score for dental
fluorosis (percent of fluorosed surfaces per subject) were computed for subpopulations and
fluoride level using fixed effects ANOVA models. The LSMeans procedure in SAS was used
to compare adjusted means for statistical significance. Bonferroni corrections were used to
limit the over all experimental type I error rate at between 5% and 10%. Statistical
significance was set at p=0.002. Inter-examiner agreements for TSIF were 79.4 % and 87.5
%; corresponding to kappa values of 0.54 (moderate agreement) and 0.64 (substantial
agreement).
TSIF scores for the 8-10 yr old children examined in 1990 are shown in Table 3 (copied
directly from Selwitz et al., 1995), with comparisons to the children examined in 1980 and
1985.
T/ Bt E i
Cur.p*i JWR o* 7SIJ" Scores »«u, \Ieaw Vfc^ 1 FJyar^J S*i i^ces o? ChiMr5 fe) 2 14 d "** b >! . .8* "-«2 —
2,\,w*i' ,-i HI isr Ti'i i> i1* •>" r- f?? '73. ^(K\
X-.pU-uil V. 2.. 7 *•» X'. '. -t .>' ^.5 ~C, ," . 1*7 -1 ->U7 v2 v.V :
>^^
Cp.ti. .3t> 522 TIi1 3!* 56 ,H C I 2tU i»w —
IK pin,,, Jl \ * W "ill ,H h? 1 ^2? ^h 'iHj
.A >pu~u i; iiin ;,•» i ;i i „"•> 82 i ^2 " :!'>0 ^I\>IT
i\.in>i-u ^ Si _i. UJ »>r "~ C2 7 l 7 , *l''i»
,*^
167 81.4 14,4 2,9 1,3 0,f3* 18,6 17.8 —
2X optimal 7f> 2,071 45.0 24,7 14,2 147 1,4 55.0 55.4 <,(»14
3X optimal 69 1,984 45.3 25,1 14,5 12.2 2,9 54.7 55,2 <.OCt"4
4Xoptitst! 57 1^70 3&4 24.9 15,5 18,3 3.1 M.& S9.8 <,t»lt
^rfi.to (riie^l >r> c *n i >u6 c^
. tt>i2* -i 'Hifaf"J
^j *, i 1C^,H ^ H E'^
TSIF scores for the 14-16 yr old children examined in 1990 are shown in Table 4 (copied
directly from Selwitz et al., 1995), with comparisons to the children examined in 1980 and
1985.
239
January, 2008
-------
Other Effects
TABJj-4
CaffipjrBoi: of TSIF^wra^Mj Mean P««m o£ Fli4«o«s«?*! Surfaces *0-ChiM~ffl mAgg Group 2 foCou!ffit*"ude« with
„*,» CK C.ci S,, JvC", I 2 3 4-" F.ui-rwoJ1 VF>1
OjXiii. ! 7,1-," -A*? v 15 i" H fui •> *t Jl
Or !"„ *I > S ( 7i, H ? . n •>• « 28 .1 29.4 30,5
:\op--,. :: !,-•<:• sis K? m«j im 11 «.s 47,2
lAupi"1- *.r i 5 i.* M^i 1^3 in 2S S9.2 65,1
IXiW.'ui, 21- i,» " 111 "V8 lAd J21 Is. 77.5 77.8
o?u~,d: V' «;«>, «; ,"n 16 .*: n is.s 14.9
?Xiv .")! J b»3 *2i ::» 111 I if , i 47.8 48,9
^\i>.-:, ',1 '.1 l '. Mi 2i,i 34 U 3 it% 46.7 4S.4
;-—- r^r,i' KV.^^^^^. .
3,.-; inra Opt
-e1';/lllinois,"y8lt, l»SS,,iRC iV'*D
A*a;ri No. at "-Canes- \;ear. "uDj.ffroin
l«.» Chite.-er fnv DMFS ,SF)' Oiitr-dl ?-VJ.,\K
Optunt.. 224 35.3 2-WnAW
2XOJU-.U 100 52.0 I7u.29i 432 .001 1
SXo^tind* "132 57.6 21 2" > 577 <,Mtti
^Xop^j, 93 44.1, 2'1'\-i. 2SS ,043
Optima. 250 44.0 ZK, ( M —
2Xojwm>. 125 53.6 x. So 1.26 33.5 .003
iXof'unv. 159 541 ' ,.->iHj3, 46.6 <.£»!t
*i\rtpu™a' 91 48.4 J 91 (31i 32.0 ,012
Opijn^ 258 51,9 1S5(".3) —
?\opti.T<,' 103 58.1 I4^l2s> 21.6 .235
j\fp:injl 117 36,4 14H2") 23,8 .176
X plural 77 50,7 1^(33'. 0.0 ,989
fH,^, nrj^l,1^ ^12^ ? J wt.^u '^ '.wc^ sr Jr Jtipio cotr^ar^ijrifc ^s^i^ JieSo:^faTO'mp^3c®duFe.
The high level of cavities at optimal limit the usefulness of this data on the cavities.
Tables 3 and 4 of Selwitz et al (follow) illustrate the percent distribution of TSIF scores on
permanent tooth surfaces across all subjects, and the mean percent of fluorosed surfaces per
subject (MPFS) among optimal and above-optimal communities.
240
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
TABU '
.V,,to,F V...1 N.,..,. ', J, in^u,,.,. ^.r^.c, -So.^s .-.^aefcr
«',' rni,-i-i s,_t ,,t>.. c i J 3 -" nooriiieJ* W7FSI- D ;.'n..-> eh»t
Opu~uJ .3 iW M: ,na 2.3 1.6 0.1 18,8 18,2 —
It vt -U •> » " Sni >li 6.9 6,? 0.* 47.0 47.3 «V (
•>.,>j>l"-il W ' V ' 10.9 8.1 1.9 51,5 52.4 <• (H.T.
j\ntT«l =W '*> in, •**-! 17,1 19.7 44 69.7 69,?, < 1)4
op.au .. =* i E ~; ;i ^ 5.6 i,s 0,1 mo ie.t
IX^vi,^ -; * : H 3'i 11.6 8,7 1.3 S2.C 52.8 - IOJ
•'Xjpt-w! '•> ^'2h !i 2J 12.3 8,2 1.1 52.0 S0.9 «>,•"_
Ort.-iJ <-• ^r Si^ 4i 2.9 1,3 O.C» W.S 17,8 —
ae>F'"'! "H ^ 7 li« ."> 7 14,2 14,7 1.4 55,0 55,6 •=. It
,Xoj.i.!,[ v( .1- 4,1 :~ 14.5 12.2 H 54.7 55.2 ,.'< f
^A'*^11 J U' 41* Hi's f » h-r.v,^ ^a '4 si"\ih\^
MNT ^ L i s us F ^ 4-u ^« ife .1 * *, it! t^
TABit 4
^.*5" liH^l SH;« A.^BT"% ff-nptiHTd.1 \\ 3f SJ .Fltt0jrlQ.fi L-fiS fils, Jtii*HCt~S, l^feltr I.Ss5f STiil iV^O
A'v" ut" ** \i> ».*" ^!( LI ^ . s*1" bi" *n ix* rS-lf- Sci >'°¥iti ^ S'l*1 ''i^t*** ^-v-t11,? ff
j vn C-i &-.T, S ,-n-i ^ i i 1 4 7 ?i-wnw-i" \ ">Ft,^ Di.t iron Dpi
['>,. -ul i 'J-.ii Hs^ ^ IS 0.8 0.0 . i TI.I —
1 ...... al * 2^U - " 2: 1 7 A 5.0 0,1 ".I % 38,4 <,001|
i\ist.rrJ ^ ","11 T. ?i i 13.7 9.6 1,1 '.-.1 45,5 * 7"^ ''1 6 4* 2,8 0,1 73 > 36,3 —
Ao;>i-r. "i 1 4U2 ^i 125 18.6 13,8 1,4 Ssi SF.2 <,001t
AVI , ^ 1 !•> ' x 1H IS.2 13.6 IS ,>J 3 6ft I <,001^
<\t>\ i.«x 2r' 114^ ^^ ""H 1&8 2,1 5,8 77 1 77,8 <,OD1.t
•H
O->i -a *•"! 'i,i>>4 h;7 , ij 35 o.i IT! MJ ,_
2^-i-n «* w ,->•, •>".! 13.1 11,0 OS ,'i 48.9 <.001t
% ^ if jf- •* '"S ^^i ^^,^. f" 1 124 l*t ^ 30 ofa m 45 4 < ODl'f'
v\ c^t r-* ^ - 2^- j5 S 2, i iai) 24J 3,1 ^^ 7 67.6 <>OQ!f
"^ muf us < Et f '« U 2 i A] 41 ^«j u Lf*\ **i s. st L i}. e t-i^pp * sin8' .sing thf Bonferrctni proGasur&
Dental fluorosis seen in children in areas with optimal water fluoride levels appeared to
increase from 1980 to 1985, but did not continue to increase from 1985 to 1990. Dental
fluorosis seen in children in areas with 4X above optimal water fluoride levels remained
stable or showed no sustained increase from 1980 to 1990. Apparent increase in prevalence
of fluorosis observed for all permanent tooth surfaces (younger children in optimal
community and older children in optimal + 2X and 3X communities) declined by 1990 to
levels observed 10 years before; explanations for this shift can not be determined with
certainty.
A clear majority of tooth surfaces affected by dental fluorosis at the optimal level received a
TSIF score of 1. At above-optimal water fluoride concentrations, dental fluorosis either
remained stable or demonstrated no sustained increase over the decade-long study.
For definitions and descriptions of scales and indices, please see Section 2 and List of
Acronyms.
Driscoll, WS et al 1983. Prevalence of dental caries and dental fluorosis in areas with optimal
241
January, 2008
-------
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/dot
e
DMO
11/30/2006
and
12/15/2006
PROFILER'S ESTEM.
NOEL/NOAEL
PROFILER'S ESTEM.
LOEL/LOAEL
POTENTIAL
SUITABILITY FOR DOSE-
RESPONSE MODELING:
CRITICAL EFFECTS
and above-optimal water fluoride concentrations. J. Am. Dent. Assoc. 107: 42-47.
Concerns regarding confounding of 4X community category (fluoride concentration of
Bushnell community water supply altered dramatically in 1982), thus affecting interpretation
of 4X community findings. As in many epidemiological studies, fluoride ingestion from
alternative sources is not characterized.
Dental fluorosis observed at all fluoride concentrations examined; thus, estimating a NOAEL
is not possible from these data.
The lowest fluoride concentration at which opacities were observed is 1 ppm.
Not suitable Q, Poor (J, Medium (X), Strong (_)
Data collected for combined "4X" communities for 1990 (and possibly 1985) is compromised
by significant downward concentration of fluoride in the community of Bushnell due to 1982
installation of water softening treatment unit to community water supply. Uncompromised
dose response may be possible with optimal, 2X and 3X communities while acknowledging
that alternate sources of fluoride have not been controlled; to be determined.
Dental fluorosis and caries
242
January, 2008
-------
Stephen, K.W., L.M.D. Macpherson, W.H. Gilmour, R.A.M. Stuart and M.C.W. Merrett. 2002. A
blind caries and fluorosis prevalence study of school-children in naturally fluoridated and
nonfluoridated townships of Morayshire, Scotland. Community Dent Oral Epidemiol
30:70-9.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT
Dental caries and fluorosis
Blind prevalence study of dental caries and fluorosis.
Scotland/ Morayshire; Burghead, Findhorn, Kinlos, Buckie and Portessie. Children from
school grade 1 (aged 5-6 years) and grade 4-7 (aged 8-12 years). 70 lifetime and 31 school-
lifetime (permanently present therein since commencing full-time schooling at age 4.5/5
years) out of an eligible total of 125 children exposed to 1 ppm F in drinking water (F
subjects); and 179 lifetime and 37 school-lifetime children out of 281 eligible subjects
exposed to drinking water with 0.03 ppm F (N-F subjects).
The socioeconomic status (SES) analysis showed that 17% of F subjects were in "high" SES
groups I or II, 75% in "nonmanual" group III and 8% in "manual groups" IV or V. For the
N-F children, the corresponding percentages were 23%, 60% and 17%, thus revealing a
higher percentage of N-F subjects at either end of the SES scale.
The study population is presented in Table 1 taken directly from Stephen, 2002.
TabK'l, Distribution uf lifetime tL; and *»;hu'jl-Jifi,'timi' i'5-L;
resident Mi"*rciv* pvup-
tags
P N-F
Age IJT) L S-L L S-L
5-6 15 43
8 19 11 23 8
9 12 5 39 11
10 11 7 29 9
11-12 13 8 45 9
70 31 179 37
101 216
A total of 15 F and 43 N-F 5/6 year-old children were examined for caries. For the 8-12
year-olds, 55 life-time and 31 school-lifetime F children and 136 lifetime and 37 school-
lifetime N-F children were examined for caries; only the lifetime children in the F and N-F
groups were examined for fluorosis.
Children from non-fluoridated water communities (N-F subjects) served as controls.
Exposure from birth to age 5-6 years to fluoridated (F) water (n = 15) or nonfluoridated (N-
F) water (n= 43). Exposure from birth to age 8-12 (n= 55 for F and n= 136 for N-F) or
only during school-lifetime beginning at age 4.5-5 yr (3 1 F; 37 N-F).
The drinking water in Burghead, Findhorn, and Kinlos was naturally fluoridated at a level of
1 ppm. The drinking water in Buckie and Portessie was non-fluoridated (N-F) with a
fluoride concentration of 0.03 ppm.
A simple parental questionnaire assessment was circulated along with the participation
permission slip to determine if and when any regular fluoride supplement and dentifrice
243
January, 2008
-------
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
usage had occurred from 0-6, 7-1 1 and 12-23 months of age and from 2-3 years or >3 years
of age; and also to attempt to determine the brand(s) of fluoride/nonfluoride dentifrice
currently and previously used by each subject.
None provided.
The objective of the study was to determine the prevalence of dental caries and fluorosis in
Grade 1 (aged 5/6 year) and Grade 4-7 (aged 8-12 years) children from three naturally
water-fluoridated (Ippm) communities and two nearby nonfluoridated communities (0.03
ppm F) in rural Moray shire, Scotland. A blind clinical caries study of 5-6 year-old lifetime
and 8-12 year-old lifetime/schooltime residents was conducted. In addition, 8-12 year-old
lifetime residents of the fluoridated or nonfluoridated communities were examined for
dental fluorosis of the permanent teeth. To ensure subject anonymity, children were
requested to attend the examination without any obvious school-identifying and hence
community-identifying apparel. Each child was asked about their own perception of the
aesthetics of their maxillary front teeth. Fluorosis was assessed clinically using the
Thylstrup-Fejerskov Index (TFI, see NRC, 2006, pages 88-89), as well as photographically.
The photographic slides were later blindly scored by four dental and two lay jurors,
alongside the UK benchmark mildly mottled (TFI=2) fluorosis comparator slide, judged in
previous studies to be aesthetically lay-acceptable. Information on their child's fluoride
supplement and dentifrice usage histories was obtained through parental questionnaires.
A head-of-household occupation information was requested to enable socioeconomic status
comparisons to be made as per Registrar General for Scotland classification criteria
(Anonymous, 1992).
DMFT/s and DMFT/S assessments (with random 10% re-examination) were carried out by
the lead study author. Subjects were examined supine as per the 1994-modified Scottish
Health Boards' Dental Epidemiological Programme (SHBDEP), at the "Dentinal 2V" level
of caries detection (i.e., " . . .definite dentinal caries evidence by visual inspection, even in the
absence of clinical cavitation"), supplemented by the use of a ball-ended, disposable CPITN
periodontal probe. As per SHBDEP acceptable practice, no sharp explorer or air-drying was
used, but gauze was available for tooth-surface cleansing where required.
For fluorosis scoring, a lay assistant asked each 8-12 year-old lifetime subject, the standard
SHBDEP question: "Are you aware of any marks on you upper front teeth which will not
brush off?". Clinically, the labial surface of teeth 13-23 were assessed, without drying, for
diffuse, homologous tooth mottling as per the TFI criteria (see NRC, 2006, pages 88-89)
with 10% re-examination. Color positive transparency photographs were obtained of these
teeth using a Yashica "Dental Eye" camera. The slides were viewed "blind" and scored
randomly under standardized projection conditions by two study authors, by two other
dental and two lay staff "jury" members with a 10% random re-viewing for inter- and intra-
observer agreement calculations. The slides were projected on a screen. Simultaneously, on
a separate, identical screen, a single "Fluorosis Impact Factor" color slide (TFI=2) was
projected. This level of diffuse, symmetrical mottling has been established as the most
"aesthetically acceptable" to 85% of 534 English teenagers shown a series of colored
photographs of non-mottled and symmetrically diffuse mottled (TFI=0-4) maxillary front
teeth (Hawley, 1996; see NRC, 2006, page 98). Panelists were given a brief "slide tutorial"
to remind them of the photographically visible mottling criteria appropriate to award a
diagnosis of fluorosis from TFI=1 and upwards.
Caries data were compared between groups by the Mann- Whitney U-test, with the Chi-
Square test applied to compare percentages between groups. Cohen's Kappa was used to
measure intra-and inter-observer agreements for categorical variables, while Dahlberg's
Direct Error Variance Method was employed to assess the intra-observer agreement
(reliability coefficient) for clinically scored caries and diffuse, symmetrical, dental fluorosis
data.
Study results in Tables 3 and 4 are shown directly from Stephen, 2002.
244
January, 2008
-------
TableB. Mean |SD) dental earfe* data for water-fluoridated (F) and naafltumdated (N-B) Moraysliire papils (Mantv-WhJtaey
U)
.Iffltt
dm-fe
5-6 vr
P; 15
N-P: 43
Syr
9jt
10 yi
11-12 yr
P: 30
N-P: 31
F; 17
N-P: 50
P: 18
N-F: 38
P:21
375%
i" ' 4
i S4 il SM
S'" (1 SW|
Till* 4 f.iunK'j 'i- U^ Hi -^i*ii^l 'Jim iJU u TFf
jtnnOii'itii1db-»1 iN-Pi j.upil h^tiil pi i tipt-^i\'U^ t^
ni tit"
lnTAii
irtii'1 f.n i. ^tst.itrei "f IjMini* tJunn*1a^Gd (F) ami
^ •.yiiiljt dat.i hit EfH TPI ^ ^t ^
N-F
Age (yr)
» (TFI teeth)
I ' Ntll'|M t<)
n (TFI teeth)
8
9
10
11-12
TbtM
TF!>2
14
12
12
16
12
16
12
24
P-0,045
P - 0.25
STUDY AUTHORS'
CONCLUSIONS:
The study authors concluded that considerable caries benefit has accrued to those
Morayshire rural children who have received naturally fluoridated water (at 1 ppm)
throughout their lives, as compared to their socioeconomically similar, nonfluoridated rural
counterparts. Only borderline mild fluorosis disadvantages were noted clinically and none
by the subjects' own aesthetic perceptions. No evidence was found to suggest any delay in
permanent tooth eruption patterns of subjects in the fluoridated water group.
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
Anonymous. 1992. Definitions of social class and socio-economic groups.
Edinburgh:General Register Office (Scotland).
PROFILER'S
REMARKS
Initials/Date
VAD/01-05-
07
The children in the study were from a rural area and were most likely all white (not stated in
the study report), and therefore, the population was not representative of the U.S. general
population. The number of subjects in some categories was small and there were unequal
numbers for some comparisons of fluoridated and nonfluoridated communities. For
example, there were 15 children from fluoridated communities and 43 from nonfluoridated
communities in the 5-6 year-old group. Only 4 children in the fluoridated and 4 children in
the nonfluoridated water groups had TFI scores of greater than 2 on a scale of 1-8 (see NRC,
2006, page 89); therefore, results are applicable only to the mild fluorosis category.
Data not useful for evaluating the occurrence of severe fluorosis. Some information is
provided that might be used in the analysis of relative source contributions.
PROFILER'S ESTEM.
NOAEL
The study design did not estimate a NOAEL.
245
January, 2008
-------
PROFILER'S ESTIM.
LOAEL
SUITABILITY FOR DOSE
RESPONSE MODELING
CRITICAL EFFECTS:
The study design did not estimate a LOAEL.
Not suitable,O; Poor (X_); Medium (); Strong (_)
Only two exposure levels were evaluated and only
mild levels of fluorosis were recorded.
Dental caries (deciduous and permanent teeth) and fluorosis (permanent teeth 13-23)
246
January, 2008
-------
Striffler, DF. 1955. Fluoridation in New Mexico: It's Present Status. NM State Dent J 5(2):3-l 1.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
POPULATION STUDIED:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
Dental caries, fluorosis, gingivitis.
Cohort
263 junior-high school age children residing in Lordsburg, New Mexico (NM) where the drinking
water was 3.25 ppm fluoride.
573 junior-high school age children residing in Belen, NM where the drinking water was 0.9 ppm
fluoride.
485 junior-high school age children residing in Lovington, NM where the drinking water was 0.8
ppm fluoride.
888 junior-high school age children residing in Santa Fe, NM where the drinking water had only
traces of fluoride.
Water history was assessed for children from birth until 8 years of age.
2290 New Mexico junior-high-age children were surveyed. The subjects were grouped based on
residence in four communities with various levels of fluoride in the drinking water as follows : 263
subjects in Lordsburg (3.25 ppm F in the drinking water), 573 in Belen (0.9 ppm F), 485 in Lovington
(0.8 ppm F), and 888 in Santa Fe (traces of F).
In each community, the students were surveyed for drinking water history and their parents were
surveyed to verify the information. Subjects were inspected for caries experience, for gingivitis, and
for chronic endemic dental fluorosis.
All fluoride determinations were made by the NM Department of Public Health Laboratory,
Chemistry Section. In each case, a minimum of 12 samples were analyzed. No significant changes
had been made in the water supplies since before the subjects studied had been born.
2290 NM junior-high-age children were surveyed by the Division of Dental Health of the NM
Department of Public Health during 1954. The subjects were grouped based on residence in one of
four communities with various drinking water fluoride levels: Lordsburg (3 .25 ppm F), Belen (0.9
ppm F), Lovington (0.8 ppm F), or Santa Fe (traces of F). This age group was selected because it is
the earliest that the effects of fluoride could be noted in all of the permanent teeth (excluding third
molars).
In each community, the students were checked for drinking water history in four ways. 1) Each
student filled out a survey which asked questions regarding place of birth, continuous residence in the
community in which they currently resided, and whether they had been away from that place for >90
days at any one time from birth through 8 years of age. 2) Parents completed an identical survey. 3)
Each subject was questioned about water history at the time of examination. 4) Any discrepancies
were resolved by a home visit by the school nurse. All children who had not had city water for any
period longer than 90 days between their birth and 8 years of age or whose water histories could not
be resolved were ruled out and not classed as continuous residents. Careful checks were made into
possible histories of drinking bottled water, ditch water, or private well water. Confounding factors
were considered: socio-economic level; food habits; amount of dental care; national origin; different
examiners/ examining conditions.
Subjects were inspected for caries experience as measured by the DMF rate, for gingivitis as
measured by the PMA index, and for chronic endemic dental fluorosis according to Dean's modified
classification. Dean's Index of Dental Fluorosis was utilized to evaluate the significance of fluorosis
in a community. The subjects were examined with new sharp No. 5 explorers, new mouth mirrors,
and Burton EENT spotlights. Compressed air was available. Bite-wing x-rays were not used.
247
January, 2008
-------
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Caries
PROFILER'S NOTE: Bite-wing x-rays were not used; however, the authors state that as long as the
clinical inspection is done is the same manner by the same examiner, as they were in this survey, then
the findings are reliably comparable.
Subjects were inspected for caries experience as measured by the DMF rate (decayed, filled, missing),
for gingivitis as measured by the PMA index (papillary, marginal, attached), and for chronic endemic
dental fluorosis according to Dean's modified classification (normal, questionable, very mild,
moderate, and severe). Dean's Index of Dental Fluorosis was utilized to evaluate the significance of
fluorosis in a community. The significance of the community fluorosis index is as follows:
Index Range Classification Remarks
00-0.4. . . Negative Indexes of little or no public health
04*06 Borderline concern as to the development of en-
demic dental fluorosis; highly impor-
tant in dental caries control,
0.6-1.0 Slight
I 0-20 . . Medium Removal of excessive fluorides from
2.0-3,0. Marked the water is recommended.
3 0-4 0 Very marked
Statistical methods were not reported.
The following table was copied directly from Striffler (1955) and shows DMF rates for each
community, broken down by total average and continuous resident average (xx indicates insufficient
number of continuous residents to warrant inclusion). Santa Fe (traces of F) continuous residents had
almost four times as much tooth decay as Belen (0.9 ppm F) continuous residents. Within Santa Fe,
continuous residents had a DMF rate of 7.3, all Santa Fe subjects had a DMF rate of 5.9 regardless of
water history, and all subjects who had been off Santa Fe water for >90 days had a DMF rate of 5.3.
Thus, dentally speaking, it was a handicap to have been born and reared on Santa Fe's F-deficient
water and an advantage to have been away from Santa Fe, even for as little as 90 days of more.
No, Con- Average
Total No. Average timiems Contmyoos
StiKlciHs Overall Residents Resident
City , Examined DMF Rate Examined DMF Rate jipro FV
Lordsburg 263 16 92 15 3 2S
Belen ,, 573 2.5 126 1.9 09
Lovington . • - 485 2 6 xx xx 0 8
Santa Fe H88 5 9 255 7 3 traces
Figure 2 was copied directly from Striffler (1955) shows the average number of decayed, missing, and
filled teeth per continuous resident by age in the four communities.
248
January, 2008
-------
Fluorosis
Gingivitis
Confounding factors
AVERAGE NUMBER
PERMANENT TEETH
LORDSBURG, BELE
XL
1
K
1
£ 3
u. £
^S '
o
A
1
<
of DeCAYED.MiSSlN6.sraf FILLED
ptr CONTINUOUS RESIDENT
BY ASE
N, ond SANTA FE, NEW MEXICO
./*«j
1
M»
ped sharply as
id severe the D
ivas present.
md with cosme
s, 32 were ques
jlic health sign
ored" section f
(1955).
re«,««y
32
30
26
0
0
= 159 g (h
0.62
; is point at which tooth decay
slight evidence of fluorosis
MF rates started to climb back
tically objectionable fluorosis.
stionable and 71 normal. The
ificance of this index is
or significance of index). The
fxw
0
16
JO
52
0
0
v) =98
No association could be made between prevalence of gingivitis and fluoride content of the water
supply. Subjects in Lordsburg with 3 .25 ppm F in the drinking water had no more gingivitis than
those in Santa Fe with only traces of F in the drinking water.
No significant differences in socio-ecomonic level were found between any of the communities,
based on percentage of families earning over $2,000 per year. An analysis of a food habits survey
conducted jointly by the NM Department of Public Health and NM A&M College showed no
appreciable differences in the consumption of milk, vegetables, proteins, or other essential nutrients.
In only one respect was there a major difference. Santa Fe children consumed significantly less
sweets than children in the other communities. In terms of dental care, Santa Fe residents had received
249
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/date
SJG/
10/18/07
PROFILER'S ESTEM.
NOEL/NOAEL
PROFILER'S ESTEM.
LOEL/ LOAEL
the most dental attention (F of the DMF rate) and had access to more dentists and public dental
clinics. No significant differences were founding national origin or racial types. All inspections were
done with same type of mouth mirror, explorer, lights, and portable chairs and were done under
similar conditions and the same examiner (with the exception of Lordsburg where two additional
dentists helped examine, but with calibration beforehand). Based on these circumstances, fluoride
content of the drinking water remains the one significant variable responsible for the extreme
differences in amounts of tooth decay.
The usual amount of fluoride recommended in most parts of the United Sates, 1.0 ppm, is too much
for the climate of most of NM; hence, 0.7 ppm is recommended as the optimum level forNM (New
Mexico department of Public Health).
The communities surveyed having adequate amounts of fluoride had as much as 70% less permanent
tooth decay amongst their continuous residents sampled than did Santa Fe with only traces of fluoride.
Other factors than fluoride such as socio-economic status, food habits, amount of dental care,
availability of dental clinics, incomparability of inspections, and racial origins were ruled out as
possible variables which might have influenced the extreme differences in decay experience amounts
in the 2,290 children examined.
No association could be made between fluoride in the drinking water and prevalence of gingivitis.
The DMF index was related to fluorosis classification increasing from a low level with mild fluorosis
to a higher level with severe fluorosis.
None.
The study was well-conducted and had adequate study design. However, the study was very poorly
designed for development of a dose response to fluoride as limited data was presented and no
statistical analyses were conducted. Junior-high school age children residing in four communities
with different fluoride levels in the drinking water were inspected for dental caries, fluorosis, and
gingivitis. This age was adequate to assess effects on permanent teeth, particularly when considering
drinking water history from birth until age 8 years old. Several variables were considered to support
the role of fluoride on the differences in DMF rates and fluorosis. The only differences in variables
found were in Santa Fe where children ate fewer sweets and had more access to dental care, factors
that would be expected to decrease the risk of caries.
Overall DMF rate decreased as community water fluoride level increased, from 5.9, 2.6, 2.5 and 1.6
in communities with traces of fluoride, 0.8, 0.9, and3.25 ppmF, respectively. The same trend was
seen in continuous residents where rates ranged from 7.3 to 1.5 in Santa Fe (traces) and Lordsburg
(3.25 ppm F), indicating that children are protected from dental caries as fluoride levels increase. No
association could be made between fluorides in the drinking water and prevalence of gingivitis.
Although DMF rates were related to fluorosis classification, with the authors claiming decreased
DMF rates as slight evidence of fluorosis was detected and increased DMF rates as fluorosis became
more severe, no data was presented to support this finding. The only data presented for fluorosis was
the community index from one community (Helen, 0.9 ppmF, annual mean temperature 56.6°F),
0.62, indicating that 0.9 ppm F may be borderline too excessive for optimal dental health (i.e.,
preventing caries while not contributing to fluorosis. This is in line with the recommendation by the
New Mexico Department of Health that 0.7 ppm F be considered optimal for most of the NM climate.
Study design was not suitable for development of a NOAEL.
Study design was suitable for development of a LOAEL for dental caries and fluorosis. Exposure to
>0.8 ppm F in the drinking water appeared to offer protective benefits with respect to dental caries
while 0.9 ppm F resulted in a community fluorosis index of 0.62, slightly above the index range
250
January, 2008
-------
POTENTIAL
SUITABILITY FOR DOSE-
RESPONSE MODELING:
CRITICAL EFFECT(S):
where removal of excessive fluorides from the water is recommended due to questionable to very
mild fluorosis.
Not suitable (), Poor (X), Medium (), Strong ()
While the study was well-conducted, the study design was poorly conducive to provide data for a
dose-response. The study indicated protective effects of fluoride in the drinking water above trace
levels (>0.8 ppm F) with respect to dental caries and questionable to very mild fluorosis at 0.9 ppm F
in only one community (no data for the other communities). Similar fluorosis data for the other
communities with various water fluoride levels were not presented, so it is unclear whether a lower
effect level or a dose-response effect would be found.
Caries, fluorosis
251
January, 2008
-------
Susheela, A.K. and M. Bhatnagar. 2002. Reversal of fluoride induced cell injury through
elimination of fluoride and consumption of diet rich in essential nutrients and antioxidants.
Molec. Cell Biochem. 234/235: 335-340.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN:
Dental and skeletal fluorosis; fluoride in serum, urine, and drinking water, and health
symptoms of people with fluorosis.
Prospective cohort
India/New Delhi and neighboring states: 10 people (6 males, 4 females, aged 8-60) with
clinical manifestations of fluorosis, who lived in rural areas.
None
Unknown
10 people who were exposed to excessively high levels of fluoride in their drinking water
and/or in their food, which resulted in their clinical diagnosis of fluorosis.
Fluoride levels in the blood, urine, and drinking water were measured using an ion
selective electrode. Exposure prior to the study initiation was not quantified, but was
confirmed by establishing that the subjects' drinking water had high fluoride levels, and
by evaluating tooth discoloration in children of the family, joint stiffness, and finding a
family history of gastrointestinal (GI) complaints that would disappear 10-15 days after
switching to safe low-fluoride water.
During the one-year intervention program, the subject's clinical symptoms and the
fluoride levels in the drinking water, blood, and urine were monitored and reported at 1-3
unspecified time points (impact assessments).
The only information provided regarding other possible sources of fluoride exposure was
that three of the patients (who had relatively low fluoride in their drinking water) ingested
food contaminated with fluoride.
Fluoride levels in the serum, urine, and drinking water were measured using ion selective
electrode technology.
Ten subjects with clinical manifestations of fluorosis were referred to the study
investigators by clinicians from hospitals in New Delhi, India, and from neighboring
states. The clinical diagnosis of fluorosis was made in hospitals on the basis of the
people's case histories, clinical complaints, forearm X-rays, and by testing fluoride levels
in their blood, urine, and drinking water. In rural areas without diagnostic facilities,
fluorosis was diagnosed after first determining that the drinking water had high fluoride
levels. Then the following were evaluated: tooth discoloration of children in the family,
joint stiffness by three physical tests in the subject (ability to bend over and touch the toes
without bending the knees; to touch the chest with the chin; and to touch the back of the
head with the hands), and a family history of GI complaints, which would disappear 10-15
days after switching to safe water.
Once fluorosis was confirmed, the subjects participated in an intervention protocol, which
consisted of drinking safe defluoridated water from village sources or home filtration with
activated alumina, and nutritional counseling to avoid high-fluoride foods and to consume
adequate vitamins C, E, and other antioxidants. Subjects were monitored for up to a year
afterwards at three unspecified intervals (i.e., impact assessments), at which time their
serum, urine, and health status were assessed. Evaluated health manifestations included
GI complaints, muscular weakness, polyurea, polydypsea, and pain and rigidity in the
joints). A single value was provided for the water fluoride concentration during
252
January, 2008
-------
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Fluoride levels in the
drinking water, serum, and
urine of fluorosis patients
Health symptoms of
fluorosis patients
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date
REMARKS SM/1/10/07
intervention, with no description of how/when the value was obtained.
Subjects were monitored for levels of fluoride in serum, urine, and drinking water, and
health symptoms on 1-3 occasions for up to a year after the beginning of fluoride
intervention.
No statistical analysis was conducted.
In all subjects, serum and urine fluoride levels progressively decreased over the course of
the one-year intervention period, as shown in Table 1. For 2/10 of the subjects (1 and 2),
serum fluoride was reduced to levels considered normal (0.02 mg/L). Urinary fluoride
levels were still above those considered normal (0. 1 mg/L) for all subjects by the third
(last) impact assessment. Water fluoride concentration during the intervention period was
significantly lower than prior to intervention for 7 of the 10 subjects, and was unchanged
for the remaining three subjects, who ate food contaminated with fluoride.
Fs^fe Flu rid \<* Utiitafrnt 'itiifhi twis before and duriog tntCTvratioH
Patientn* Himril in Jrmhng 'frter imi_ ti Fiui id«~ n M.™ Hmi. Ji l-tu ind™ in urm |mc 1
I* lure lukr i"iiu n Puiia^nl i lit! ii fa tua uiti r cnti n I'uruumt nliu i ButfiR i ikr i"nlm During i i r \.n I n
i"[. :• IA *i\ ]-i 2 ' "" 11
1, "Uin I _~ y ilr U lA i ' 1 u_ • Q< 4 " 1 n i n
2. * w i "ii n i; n in n' ji.; HIM | •< .i 11:1
3, 2 U~ ' ' i'22 I]' '« l> ^ _4 HI ]M!" « 'i >
4. I "4 1 »<>•, i 4 "I HI' ;_| I if «l II i|
5, : Min . ail m . Mil In tin n» 41! '1 II Ml
6.* t Ui Uii II Jj 1 In 311 jn 2 i [ 4t *l M ™U
7,« i (s n i lim) 14 1 fin t,iin
8. ;nn I ^ inH l<4 ^Oi i si*
9.* i' 14 i 14 IM i M4 ii "i •* >l
10. nun n s. n in 114 i .- i mi
Pemui^ibii liiait Htim id" m f}n hug ^ it^r 1 'Img 1 trie *"" rmihipp r linnlut tlwu dt m ^rtin l'i'2i it. L |c^] N nnal upc itmit J lluiindi in
urine;! lmtl[M]I> Imp it i s> mi-ut '1 uj uumiiunil »nh llonntl-
All 10 patients had complete recovery of their health symptoms by the end of the third
impact assessment, as shown in Table 2. Recovery was the quickest for GI complaints,
with 70% of the participants reporting a recovery at the first impact assessment.
Symptoms were ameliorated more quickly in subjects who drank low-fluoride water and
had nutritional supplements, as compared to those who only drank the low-fluoride water.
"b c "" HI iillh tntpr 11 fat* c\j.s * 1 h tht-pjranc m = UM
Maiiii1* l-ilmn., pi r at ittln ti u P^R "nta JL ro \ °r Juitnc i ti n LW\ >ti
Ixti^a mkr toll n Pimpuita id^tr int tiiul -vrnplunf !'«( ~« ]'>M
IIu lilar^.ilns.^, Hi 4* n i i^kt-nx ^r
Pfl un J ^U 2* 'if i mjl t IM u
Fut ^p-*.i U 21 4fi I nipli t r^ r
fdm in1neidiH tuth I'Hnt *n t 'il < ^-ilpltE n> •". r
Susheela and Bhatnagar (2002) concluded that fluorosis can be reversed. Removing
fluoride sources and a diet containing essential nutrients and antioxidants can significantly
improve health (i.e. reduce fluoride toxicity) and reduce fluoride in the urine and serum of
fluorosis patients. This was shown in 10 patients who had complete recovery of a variety
of clinical symptoms and lower urine and serum fluorine after reducing their intake of
fluoride in the drinking water.
The study unambiguously showed that reducing water fluoride intake led to decreased
fluoride levels in the serum and urine of fluorosis patients, as well as recovery from a
number of health symptoms that appeared to be fluorosis-induced.
253
January, 2008
-------
PROFILER'S ESTEM.
NOEL/NOAEL
PROFILER'S ESTEM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
The data may be useful for estimating the levels of serum fluoride associated with adverse
health effects.
Insufficient data were provided, however, for a quantitative dose-response assessment of
water fluoride levels and fluorosis in the subjects, or of the decrease of urinary and serum
fluoride with time. For example, there were no quantitative estimates of the cumulative
fluoride intake of the 10 subjects, and the time at which the serum and urine were
collected were not provided. Also, the study had no reference control group.
Cannot be determined from this study.
Cannot be determined from this study.
Not suitable (x), Poor ( ), Medium ( ), Strong ( )
Data were insufficient for a quantitative dose-response assessment of water fluoride levels
and fluorosis, or for the decrease of urinary and serum fluoride with time. No reference
control group was provided.
Increased serum and urinary fluoride levels, associated with adverse health symptoms (GI
complaints, muscular weakness, polyurea, polydypsea, and pain and rigidity in the joints).
254
January, 2008
-------
Szpunar, S.M. and B.A. Burt. 1988. Dental caries, fluorosis, and fluoride exposure in Michigan
schoolchildren. J Dent Res 67(5):802-806.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT
ANALYTICAL METHODS:
STUDY DESIGN
Dental caries and fluorosis (permanent teeth)
Case sectional survey
U.S./Michigan: Study participation was sought from children aged 6-12 years who were
continuous residents of four Michigan communities with varying levels of fluoride in the
drinking water. Of the 1 103 children who returned a questionnaire, 556 were continuous
residents of the following communities and qualified for study participation: Cadillac
(n=13 1), Hudson (n=133), Redford (249) and Richmond (n=43). Table 1 is included below
as copied directly from Szpunar and Burt, 1988.
TABLE 1
OF BY AGE,
AND
Aje in Yem
Community Gender N Unto 6 6-7 8-f 10-11 12 +
CwHllw M 56 I 21 13 It 3
F 75 6 IS 20 25 6
M 60 7 11 17 17 0
F U 12 25 If 11 4
Redforf M 12? 6 M 39 19 1
F 122 11 45 33 2? §
M II 0 5 9 4 i
¥ 25 0 10 7 11
None.
Children were lifelong residents of the communities so their exposure period was birth to 6-
12 years old. The date when the study took place was not provided.
Children aged 6-12 years were exposed to drinking water with one of four fluoride levels:
Cadillac (0.0 ppm), Hudson (0.8 ppm), Redford (1.0 ppm) or Richmond (1.2 ppm).
Only fluoride concentration in water was evaluated. Average water fluoride concentrations
were obtained from the Michigan Department of Public Health's listings of the fluoridation
status of communities served by public water supplies for the past 20 years. The levels for
the Cadillac, Hudson, Redford and Richmond communities were 0.0, 0.8, 1.0 and 1.2 ppm,
respectively.
Not provided.
The objective of the study was to investigate the prevalence of dental caries and fluorosis in
the permanent teeth of Michigan schoolchildren, residing in four different communities, to
the various concentrations of fluoride in the water supplies. Questionnaires on demographic
information, residence history, details of fluoride exposure and use of dental services and
infant nutrition were completed by parents. A total of 556 children who were continuous
residents were examined for dental caries by means of the DMFS (decayed, missing or
filled surfaces) index applied with the criteria of the National Institute of Dental Research
(MDR). Following the caries examination, the Tooth Surface Index of Fluorosis (TSIF; see
255
January, 2008
-------
NRC, 2006, page 90) was used to measure the prevalence and severity of dental fluorosis.
Russell's criteria for differential diagnosis of fluorosis were also employed. (Russell, 1961).
Equipment used in the screening included a portable dental chair, Rolux fiber-optic light,
mouth mirrors and no. 23 explorers. One examiner conducted all the examinations. A
reproducibility test was conducted on 24 randomly selected subjects to test for intra-
examiner reliability.
PARAMETERS
MONITORED:
The permanent teeth of children were examined for dental caries by means of the DMFS
index applied with the criteria of the NIDR. Following the caries examination, the TSIF was
used to measure the prevalence and severity of dental fluorosis. Russell's criteria for
differential diagnosis of fluorosis were also employed.
PROFILER'S NOTE: The TSIF values are described in Section 2.
STATISTICAL METHODS:
All statistical analyses were carried out by means of the Michigan Interactive Data Analysis
System. The analysis of variance (ANOVA) was used to test differences in mean DMFS
values among different age groups, by area of residence, and by use of fluoride sources and
dental services. Multiple pair-wise comparisons were performed when indicated by
ANOVA results. Categorical methods were used to determine a list of variables that were
consistently associated with caries or fluorosis. Then, these variables, along with the age
and education variables, were used as the independent predictors in logistic regression
analysis. The intra-examiner reliability was tested on 24 randomly-selected children. The
percentage agreement between DMFS scores from the first and second examinations was
calculated and Pearson product-moment correlations were computed for the first and second
examination results for both caries and fluorosis. The kappa statistic was used to quantify
agreement in scoring the presence or absence of fluorosis between the two examinations.
The agreement between the two examinations for DMFS scores was 96%. The Pearson
product-moment correlation between DMFS scores from the first and second examinations
was 0.92 (pO.Ol). The agreement between first and second examinations for fluorosis
(presence/absence) was 92%. For these data, the kappa value was 0.85, suggesting a high
degree of consistency. The Pearson product-moment correlation for the sum of permanent
surfaces scored as free from fluorosis was 0.94 (O.01).
RESULTS:
Study results in Tables 2-5 are shown directly from Szpunar and Burt, 1988.
I
OF WHO WERE
AND AND OF
BY
ftercent
Caries- It «*
Meaa
DMHT
(0.0
Hudson
(QJ& ppm)
55.1 132 IM 12,3
SiJ 1,04* l.$4» 31.6
73,1 0.611 IKSP 49,0
ft.1 0.58* §,»« 51.2
— .. ^ ^§^L=^^ iSL .... , t-jJL.^, _^ lgj_..
'XJ =• 14.713, df - 2, p<0.002.
*nMFT from p « 0,0197; DMFS, p * 0.0215,
3DMFF and DMFS differed From Carfiilic, p «=
*DMFT diflftrem from Csdtilic, p - 0.0073; DMFS, p *= 0.0045-
Y - 55,594, if *. 3, p <0.0001.
fUchncnd
(1.2 ppm)
All
with
256
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
TABU 3
OF BY AGE, AND
OF ALL
BY TSIF
Age % pTev, TSIF Score
fYft.) N R«f tMis' 1 2 1-1
<§ SO 2,0 100.0 00
6-7 204 31.9 m.5 1.5
§-9 157 49,7 95.8 4,2
10-11 123 42.3 WO 1.0
12* 22 21-3 §.0 -
'X3 - 41.995, 41 » 4, p
-------
FOUND IN NRC (2006)
PROFILER'S
REMARKS
Initials/Date
VAD/03-10-
07
PROFILER'S ESTEM.
NOAEL
PROFILER'S ESTIM.
LOAEL
SUITABILITY FOR DOSE
RESPONSE MODELING
CRITICAL EFFECTS:
Although demographic information was collected, it was not provided in the study report.
Therefore, no determination can be made about whether the study population was
representative of the U.S. general population. Comparison of the questionnaire responses
from children who were examined and children who were not examined (but returned
questionnaires) showed significant differences in some characteristics of dental attendance
and oral hygiene practices, suggesting that non-participants received more frequent dental
care and practiced better oral hygiene. Various sections of the study report indicate that
study participants were 6-12 years of age. However, in Table 1, which included the number
of continuous residents by age, gender and community of residence, there were categories
for "under 6" and "12+" years. In addition, the number of participants in each group and
number in each age range of exposure level varied widely. The fluoride concentration in the
drinking water for Redford was given as " 1 .0 ppm adjusted" under the Material and
Methods section but no explanation on the type of adjustment was provided. Only mild
fluorosis was reported so results are restricted to this level of the effect.
The technical reviewer agrees that the study is a poor model for dose response modelling
because all of the groups (even exposed to water with 0.0 ppm fluoride) experienced
fluorosis, but it does demonstrate the trend of decreasing incidence of caries and increasing
incidence of fluorosis as the fluoride levels increase. The study also suggests other sources
of fluoride must be used in this group of children.
The study design did not identify a no-flurosis intake dose. Mild fluorosis of the permanent
teeth was observed in 12.2% of children in the Cadillac community where the fluoride level
in the drinking water was reported as 0.0 ppm.
The study design did not identify a LOAEL.
Not suitable,O; Poor (X_); Medium (); Strong (_)
Dental fluorosis and caries (permanent teeth)
258
January, 2008
-------
Thaper et al. 1989. Prevalence and Severity of Dental Fluorosis in Primary and Permanent Teeth at
Varying Fluoride Levels. J. Indian Soc. Prev. Dent, pp.38-42
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
Dental fluorosis
Cross-sectional study of fluorosis in primary and permanent teeth and fluoride levels in
drinking water
India/State of Rajastnan, 792 children (sex not specified), 6, 8, or 10 yrs old from 16 rural
areas. The dental hygiene habits of the children and the socio-economic status of the
parents were not evaluated. The average annual maximum temperature of the area was not
reported.
None
From birth to 6, 8 or 10 yr old. It was not stated when the individual examinations were
conducted.
Exposure groups were 1.04 ppm fluoride (mean value for 4 rural areas), 2.4 ppm (mean
value for 5 areas), 3.91 ppm (mean value for 4 areas), and 6.0 ppm (mean value for two
areas). The sources of the drinking water were open or tube wells.
Drinking water was the only exposure route evaluated
Orion ion-specific electrode was used to analyze for the fluoride content in the drinking
water
The prevalence and severity of dental fluorosis in both primary and permanent teeth was
studied in Indian children, ages 6, 8, and 10 years old. The children were examined for
dental fluorosis using Dean's Index of Fluorosis (see Section 2). 10% of the children
examined each day were re-examined. The prevalence of dental fluorosis in the primary
and permanent teeth were calculated and correlated with the levels of fluoride in the
drinking water (four exposure groups were used).
Dean's Index was used to evaluate the grade of dental fluorosis.
Not stated
See Table 2 for the prevalence of dental fluorosis in primary and permanent teeth at
varying fluoride levels, Table 3 for the severity of dental fluorosis in primary teeth at
varying fluoride levels, and Table 4 for the severity of dental fluorosis in permanent teeth at
varying fluoride levels.
259
January, 2008
-------
*.#
ft
•s
*J
p
s; .s
SS
OOO o<*"iOi> Oiri*©%o r^r^^"*-
ooo OW^^T^N ot^-r^v-j or^oo
OOOO
OOf^C^CN
3
S
260
January, 2008
-------
w
ft
X
z
>
,*
l>
I™
<
1
-S
5 » e 5 • B .*, * B ^ s »; ^
i " " s.I *» -1 ' -.,-; I • - •?.!
261
January, 2008
-------
c
s
or
a
£
It
£
'J
u.
o
<• 5
x S 8 & •» /-, 8 *> ', 'J a -•' •>"
? 7 , J' J i,
«. *( ^ g 4 f 4, s;
'
J
£ ^
if
t
£ .a-'
£ „ ,- $ ;; r ?,
STUDY AUTHORS'
CONCLUSIONS:
The prevalence of dental fluorosis in primary teeth was lower as compared to permanent
teeth at all ages and at all fluoride levels. The prevalence of dental fluorosis in both primary
and permanent teeth increased with every increase of fluoride level in drinking water. At
1A ppm, the percentage of permanent teeth having objectionable fluorosis was very low
compared to at 3.91 and 6.0 ppm.
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
None
PROFILER'
SREMARKS
Initials/date
SBG 3/27/07
This study was in India, so it would not be representative of the U.S. population.
The two major deficiencies of the study were the lack of information on fluoride exposure
262
January, 2008
-------
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
from non-drinking water sources, and the lack of statistical analysis of the data. The
higher occurrence of severe fluorosis in 8-yr-olds exposed to 3.9 1 ppm fluoride than in 6, 8
or 10-yr-olds exposed to 6.00 ppm, may reflect difference in exposure to non-drinking
water fluoride. The percentage occurrence of fluorosis was based on the number of teeth
showing the effect rather than the number of individuals; therefore the data may not be
directly comparable with other studies.
Based on the data presented in Table 4, the NOAEL for severe dental fluorosis (in
permanent teeth) (Dean's Index of 4) is 2.4 ppm
Based on the data presented in Table 4, the LOAEL for severe dental fluorosis (in
permanent teeth) (Dean's Index of 4) is 3.91 ppm.
Not suitable Q, Poor (J, Medium (X), Strong (J
Although a NOAEL and a LOAEL were identified, no statistical analyses were applied to
the data, so this limits the study's suitability for dose-response modeling.
Dental fluorosis
263
January, 2008
-------
Villa, A.E., Guerrero, S., Icaza, G., Villalobos, J., Anabalon, M. 1998. Dental fluorosis in Chilean
children: evaluation of risk factors. Community Dent Oral Epidemiol. 26: 310-315.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
Dental fluorosis
Case control study
Chile/San Felipe: Public and private school children, residing in San Felipe, Chile were
categorized in two groups according to their age when artificial water fluoridation was
introduced in 1986: Group I: born after!986 (n=68) and Group II: 16-24 months old in
1986 (n=38).
Public and private school children, residing in San Felipe, Chile: Group III: >24 months
oldin!986(n=30).
Clinical examinations were performed in July-August 1996, ten years after the addition of
fluoride to the water supply (January 1986).
Children who volunteered for the study and attended one of the five schools selected at
random were categorized into groups according to their age at the initiation of water
fluoridation (1986). The mean fluoride water concentration for San Felipe was 0.93 mg/L
(range 0.65 to 1.42 mg/L). The range of fluctuation around the mean value was ±12% in
87% of the samples.
All children were clinically examined for dental fluorosis in all fully erupted permanent
teeth. No radiographs were taken during the surveys.
The reported values come from daily samples of the waterworks facility serving San
Felipe. Data on how fluoride concentrations in the water supply were measured were not
reported. Other water quality parameter data were not included in the study report.
In the study, 136 permanent residents of the optimally fluoridated community of San
Felipe, Chile were categorized into one of three groups according to their age when water
fluoridation was introduced in 1986. Group I was born after 1986 (n=68); Group II was
16-24 months old in 1986 (n= 38); and Group III was >24 months old in 1986. The study
population included children from two private and three public basic schools. Selection
of the schools was made at random from all private schools whose pupils are considered
to have high socio-economic status and from all public schools whose pupils are of low
socio-economic status. Eligible subjects were permanent residents of San Felipe,
answered the questionnaire, had no orthodontic attachments on central maxillary incisors,
(CMI) restorations orun-erupted teeth, and were not diagnosed as 'questionable' (
fluorosis score=0.5). The subjects were clinically examined for dental fluorosis in July-
August 1996.
Examinations: Enamel fluorosis was evaluated in all fully erupted permanent teeth
according to Dean's index and Russell's criteria (Russell 1961) for differentiating fluoride
and non-fluoride opacities. Teeth were not dried before inspection and were examined
under tangential natural light. Each tooth was assigned an individual score to compare
prevalence and severity of fluorosis in each group of teeth according to a slightly
modified WHO form. Examinations were performed by one examiner; approximately
12% of children were re-examined 4 weeks later in a single-blind fashion to monitor
diagnostic standards (kappa value= 0.87).
Pre-tested questionnaires were developed for this study to collect information from
parents concerning children's nursery school attendance, residence histories, and the
frequency, extent, and duration of breast-feeding and tooth brushing practices. The
interviewer and parents were blind to the child's fluorosis status. Subjects were defined
264
January, 2008
-------
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental fluorosis
as 'cases' (score >1) or 'non-cases' (control; score=0) based on their fluorosis score in the
central maxillary incisors (CMI).
Dental fluorosis was assessed and recorded according to Dean's index. Russell's (1961)
criteria were used to differentiate fluoride and non-fluoride opacities. No radiographs
were taken during the surveys.
Data were analyzed using the SAS statistical package. Logistic regression analyses were
used to develop a model of exposures associated with very mild to moderate enamel
fluorosis. Odds ratios from the regression coefficient were used to estimate the relative
risk for each factor, adjusting for confounding variables. Ninety -five percent confidence
intervals were generated for all adjusted odds ratios. The independent variables were
breastfeeding duration and age at the onset of community water fluoridation; other
independent variables included sex, socio-economic status, nursery school attendance, age
when tooth brushing started, and tea ingestion.
Table 1 was copied directly from Villa et al. (1998) and shows CMI enamel fluorosis
status by age of subject when water fluoridation began. Groups I and II had a higher
percentage of cases than Group III (5 and 3 times more cases, respectively).
%h rr^rr .1 S'limrf r nf « .i| c&S^
A^*? ^Tvur \'^ smnnlH) wnen w.?*tr H 'rri^i'ti1 hi*^.ir ran;. iGnUjl* W ps>uf*
1
Lr'iUfi ! l.nhnrn * 31 iy i2-9'
1 ,-T.,,l- It lrj-?H 1? ;,'"• .19 11 -fr3
'"noio III \tnrc ilvn 14 ,t 17 in 10.0
I,,h,i -.1 « ••>„
Table 2 also was copied directly from Villa et al. (1998) and shows the adjusted odds
ratios with 95% confidence intervals for CMI enamel fluorosis. Group I and II children
presented a statistically significant increase (OR=20.44 and 4. 15, respectively) in the odds
of CMI fluorosis compared with Group III children.
AQ 'ni s'Hi^ v*\ i'n
U "ill"" H K'T'^hf *- AUj Htv'd ''^
tgrew-r '"'R'"' °'' tl
CKwp i L'nKm »l-4 I'KJ-Mj+l
epwpir J-JS4 415 ]I!'-I^4J
Cimipl'l __ ^"i. H-TT_£1 HO'
TECHNICAL REVIEWER'S NOTE: Table images inserted above are from best
available copies.
Exclusive breast-feeding was a statistically significant protective factor in the Group I
children with a 14% significant decrease in the likelihood of CMI fluorosis (OR=0.86; CI:
0.75-0.98). An increase in the prevalence of enamel fluorosis in the posterior teeth (first
molars excluded) of children in Groups II and III was non-significantly related to the
amount of tea ingested and to an increased daily frequency of tooth brushing when
subjects were pre-schoolers.
PROFILER'S NOTE: The profiler agrees that there was an increased percentage of cases
of CMI enamel fluorosis in Groups I and II compared to Group III; however the degree of
265
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date:
REMARKS SJG/1/15/07
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM.
LOEL/ LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
statistical significance should have been included. The profiler also agrees that the risk of
developing fluorosis differed depending on the age of the subject when water fluoridation
was implemented, and decreased with exclusive breast feeding (average breast-feeding
duration of 5.5 months).
Very mild to moderate enamel fluorosis of permanent QVH was strongly associated both
with the age of the subjects when water fluoridation began and with breast-feeding
duration for children of Group I. Children born when the drinking water was already
fluoridated (Group I) had a higher risk of CMI fluorosis (OR=20.44; 95% CI: 5.00-93.48)
compared to those older than 24 months when fluoridation was implemented (Group III)
while the group 16-24 months old when water fluoridation was initiated (Group II) had an
intermediate increased risk of fluorosis (OR= 4.15; 95% CI: 1.05-16.43). An extended
period of exclusive breast-feeding appears to be a protective factor for the Group I
children (OR=0.86; 95% CI: 0.75-0.98) but not for the Group III children, who did not
ingest fluoridated water during their first 24 months of life.
Tea ingestion and the use of fluoridated dentifrice started at similar average ages (32-35
months old). Fluoride supplement use was almost negligible. No significant differences
were observed between children with high or low socio-economic status, possibly because
there were no differences in use of reconstituted powdered milk.
The current results suggest that under Chilean conditions, the increased prevalence of
CMI fluorosis may be associated with fluoridated water (0.9 mg/L, with range of 0.65 to
1 .42 mg F/L) intake during the first 2 years of life.
CMI= central maxillary incisors
Russell, A.L. (1961). The differential diagnosis of fluoride and non-fluoride enamel
opacities. Journal of Public Health Dentistry 21, 143-146.
World Health Organization (1987). Oral Health Surveys: Basic Methods. 3rd ed. WHO,
Geneva.
The study was well-conducted and had adequate study design. However, the study was
not designed for development of a dose response to fluoride as the emphasis was on
establishing risk factors for fluorosis in children born before and after the addition of
fluoride in the water supply (1986). Further, the sample size of this study population
(n=68, 38, and 30 per group) seems relatively small.
The risk of developing CMI enamel fluorosis increased in children exposed to fluoridated
water from birth (OR=20.44) compared to those exposed after 24 months of age.
However, it is important to note that this does not indicate a cause-effect relationship.
The risks of dental fluorosis of teeth other than CMI and first molars that mineralize later
in life, and which may be associated with tea ingestion and frequent tooth brushing with
fluoridated toothpaste, could not be evaluated in the current study.
Study design was not suitable for development of a NOAEL for fluorosis.
Study design was not suitable for development of a LOAEL for fluorosis.
Not suitable (X ), Poor (_ ), Medium (J, Strong ( J
While the study was well-conducted, the study design was not conducive to provide data
for a dose-response; however, it does provide some information that will contribute to
identifying the sensitive exposure window. The study indicated odds ratios for CMI
dental fluorosis in children residing in San Felipe, Chile, dependent on age when water
266
January, 2008
-------
CRITICAL EFFECT(S):
fluoridation was implemented. The study did not address any
plaque or gingivitis.
issues of dental caries,
Prevalence of dental fluorosis in central maxillary incisors.
267
January, 2008
-------
Warnakulasuriya, K.A.A.S., S. Balasuriya, P.A.J. Perera, and L.C. Peiris. 1992. Determining
optimal levels of fluoride in drinking water for hot, dry climates - a case study in Sri
Lanka. Commun. Dent. Oral Epidemiol. 20: 364-367.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
STUDY DESIGN
Dental fluorosis, in conjunction with dental caries (DMFT: decayed, missing, and filled
permanent teeth index)
Case control, retrospective
Asia/Sri Lanka: 380 children 14 years old (191 male, 189 female) who lived in rural
districts of similar socioeconomic status in four geographic areas (Galewela,
Wariyapola, Kekirawa, and Rambukkana). The areas had similar altitudes, with annual
maximum temperature of 29-32°C.
The reference (i.e., control) groups were: (1) 211 children (of the 380 total) exposed to
<0.4 ppm in their drinking water, for evaluating caries status vs. water fluoride levels;
(2) 200 children with Dean's fluorosis index of 0 or 1, for evaluating the fluorosis
index vs. water fluoride levels; and (3) 156 children with normal teeth (fluorosis index
of 0), for evaluating the mean DMFT vs. the dental fluorosis index.
1972-1986; from birth to 14 years old
Schoolchildren aged 14 (grades 8 and 9) who lived their entire life in the four selected
geographic areas in Sri Lanka, and lived within 15 miles of their secondary school at
the time of the dental examination.
Drinking water fluoride concentrations in 380 domestic samples provided by the
children from the four geographic areas, and the community dental fluorosis index (FC1)
are shown in Table 1 .
Table i. Distribution of wafer fluoride level (ppm) in four areas
Water F^ level
loui.i.iti n tranp* Range Mean SD F.,
(.a'cwco 92 32 0,34-2.80 0.62 0.50 0.92
Wai jMx)i,t 100 32 0.09 5.60 0-61 0,70 0.89
iCtli'«*a 97 32 0.17-8.00 0.88 1.12 t.72
KJ ih^.uK 91 29 0.08-0.33 0.14 0.04 Q.1J
All 380 0.08-8.00 0.5? 0,75 0.93
* Mra . dpi >>:l maximum daily temperature in "C,
«~, t'ummii m> index of dental fluorosU.
The children had not used fluoride-containing toothpaste of any other fluoride therapies
during the time of their permanent tooth development (age not specified, presumably
up to age 6).
Water fluoride concentrations were measured using an (Orion) ion-specific electrode.
The water ionic strength was controlled with unspecified buffer of pH 5.0-5.5. All
measurements were made within 2-3 weeks of sample collection. Data were not
provided for any other water quality parameters.
The four geographic regions were selected after pilot studies to give a wide range of
ground water fluoride concentrations. Children aged 14, and who had lived their entire
life within their four geographic areas, were examined in November and December
1986 at each of their (four) schools. Teeth were examined using indirect natural
daylight as the only illumination, plane dental mirrors, and #23 explorers. The criteria
of Russell (1961) were used to distinguish between fluorosis and nonfluoride enamel
268
January, 2008
-------
PARAMETERS MONITORED:
STATISTICAL METHODS:
RESULTS:
Dental caries (DMFT)
Dental fluorosis
opacities, and the degree of fluorosis was classified by Dean's (1942) dental fluorosis
index. Fluorosis was assessed inbuccal and occlusal tooth surfaces of all teeth, and the
more severe score was assigned to a given tooth. The classification for a given person
was based on the most severe fluorosis seen for two or more teeth.
Fluorosis was determined in the children by one dentist, and another dentist examined
each child for dental caries using the DMFT index and the criteria of Radike (1972).
Missing teeth were assumed to have been extracted due to dental caries, unless
informed otherwise. Intra-examiner consistency scores was evaluated by duplicate
examination of 10% of the subjects each day, which found complete agreement in 66%
of the cases and agreement within 1 score in 92% of the cases for fluorosis, and the
mean DMFT score within 90.5% of the first examination.
Drinking water fluoride concentrations of the children were determined in 300 mL
water samples provided by the children from their domestic drinking water. If the
children had moved within their geographic areas, they were asked to provide, if
possible, a sample of their drinking water up to age 6.
Dental fluorosis was assessed using Dean's classification system and the criteria of
Russell (1961) to distinguish between fluorosis and nonfluoride enamel opacities. For
the Dean's Index, the authors used a score of 1 instead of 0.5 for questionable results,
thus the more severe categories ranged from 2-5. Each child was examined for dental
caries using the DMFT index.
Differences in the percent of children that were caries-free or were fluorosis-free at
various water fluoride levels were evaluated using the Chi square non-parametric test.
Differences in the mean DMFT index at various water fluoride levels were evaluated
by one-way ANOVA.
There were statistically significant differences in the percent children that were caries-
free, and in the mean DMFT scores, at five different fluoride levels (Table 2). The
differences were mainly due to the 0.6-0.79 ppm water fluoride group, which had the
lowest DMFT (1.91) and greatest percentage of caries-free children (37.5%).
Significant differences were not seen among the other four exposure groups. [Note that
in Table 2, the lowest fluoride concentration should be <0.4 ppm instead of O.04
ppm.]
Table 2, Distribut'on o*" caries sUiit,;. ny Hcuer ."uiini/t' levels
^.;ree* DMrr* ,M1WT
Water F~ (ppm) /, n '. Mea , SD reduction
<0.04 2. W 166 '•• ,!5 2!^
0.4-O.S9 49 ,V :D4 IK 2 V4 (4
0.60.79 32 :: 375 1 M 2^5 41
0.8-0.99 2" S 2-) (i 2 56 2 2"1 24
>1.0 <« ,5 2-5 2~4 2.V' Is
All 3SO "5 l',»n 3 it' 25"
* X« = 57,25;F1.0 ppm, 20%, 49%, and 3 1% had scores of 0-1, 2-3, and 4-5, respectively.
269
January, 2008
-------
Effect of dental fluorosis on
dental caries (DMFT)
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
table .1 F uorows index und water fidouue teve.s
H lurosLbindex J' «ve , - d- ^.ug «:it,-r (ppi i
(D AS\I < 1 39 o- 0 %i ',).(> 0 ~li o *-i, y*< ~> 1.0 Total
t-I lt& 69» 24 -W, 11 P-) 7(26) !2 20! 200(52)
2 + 3 52 '25l :OH!I If (>0» ,--52 *'i -.». 132(35)
4 + 5 3 ;6) 5 I'll 5 116 r> 22 19 ,," ' < 48(13)
Total 211 49 \J 2" hi 380
% fluorosis free* 69 49 ""-4 2h J,) 52
F0 0.58 0,93 1 2\t ^4 1 72 "0,93
* Scores shown are listed in Table 4. Perce iiaccs in xiie>u eso
* Includes questionable category. xl = 89,52, "• 1 .I'll
A comparison of the mean DMFT scores with the Dean's fluorosis index showed no
statistically significant differences in the groups, as shown in Table 4. The DMFT
scores were slightly lower in children with questionable or mild a fluorosis (index of 1
or 2), suggesting a mild protective effect against dental caries in these groups.
Ti.ble ** JhH ibu ion o'~ mcah DS3J i h\ isci>
Lf It o oss *nut s
rnii'T
i K*e\ (i Mea i ^t5
Nt^rnia (^\ i ?^ - > ~" -1 -Sj
Ui,»t o,..ibkM 1 ) 4»: JS"1 2W
»ci\ niH >%t %> 2 -^5 2 5->
V1i,d (3l -*-• ^45 2,7(>
\'jti:. iu H!* severe
5 1 48 3.3! 2.36
A! 380 3.01 2.57
0 !?-«„} ANOV\ - Fl^, = 1.96: l>>0.05.
Warnakulasuriya et al. (1992) concluded that 0.6-0.79 ppm in the drinking water
provided the maximal water protection against caries in these four regions in Sri Lanka
(43% lower DMFT than at <0.4 ppm fluoride, per Table 2). The authors assert that
drinking water guidelines in temperate climates, and those set by the World Health
Organization (1.5 ppm), were too high for developing countries with hot, dry climates
like Sri Lanka, and recommend 0.8 ppm as an upper limit for these populations.
Increasing the water fluoride level was associated with an increased prevalence and
severity of fluorosis, but the dose-response relationship had no threshold since fluorosis
was seen in all groups of children. Fluorosis was prevalent even at the lowest water
fluoride concentration, i.e. O.39 ppm, as 3 1% of the children had definite fluorosis
(Dean's index of 2-5). At the optimal water fluoride concentration of 0.6-0.79 ppm,
only 34% of the children were fluorosis-free (Dean's index of 0-1). The study authors
speculate that the high fluorosis index may be in part due to the high consumption of
locally grown tea rich in fluoride, or that there may be differences in the nutritional
status of the children that could influence fluoride metabolism.
The authors note that the study results may have been influenced by the mobility of
children within a given geographic region, and by the consumption of water from
schools, because wells within 15 miles of each other sometimes had 10-fold different
water fluoride concentrations.
Radike, A.W. 1972. Criteria for diagnosis of dental caries. In: Proceedings of the
Conference on the Clinical Testing of Cariostatic Agents. Chicago, American Dental
Association, 87-88.
Russell, A.L. 1961. The differential diagnosis of fluoride and nonfluoride opacities. J.
Pub. Health Dent. 21: 143-146.
270
January, 2008
-------
PROFILER'S
REMARKS
Initials/date
SM/1/09/07
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM. LOEL/
LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
A significant finding of this study is that fluoride drinking water guidelines developed
for temperate climates in Western countries may not be appropriate for populations
from developing countries in hot, dry climates. The study authors did not discuss the
reason for these differences, but presumably these may include the fact that these
people may drink more water and have dietary differences that significantly impact
their total fluoride intake.
A definitive relationship could not be drawn between the severity of dental fluorosis
and dental caries (DMFT score), as Table 4 showed there was no statistical difference
between the fluorosis index and the mean DMFT.
Two major methodological shortcoming of the study were (1) the water fluoride
concentrations were disproportionately distributed into three very narrow ranges (0.4-
0.59, 0.6-0.79, and 0.8-0.99 ppm) and two very broad ranges (<0.4 and >1.0 ppm); the
high concentration group, in particular, should have been subdivided into 2 or more
groups, and (2) there were too many uncertainties regarding the children's fluoride
exposure (potential confounders included drinking tea high in fluoride, differences in
nutritional status, and one person obtaining water from sources with vastly different
water fluoride concentrations), thus precluding the ability to establish a dose-response
for fluoride exposure vs. fluorosis.
A NOAEL for fluorosis could not be established, as even the lowest water fluoride
concentration was associated with fluorosis in 34% of the subjects.
A LOAEL was not identified for fluorosis (or for dental caries due to fluorosis).
Not suitable ( ), Poor ( ), Medium (X ), Strong ( )
Increased water fluoride concentration was correlated with increased dental fluorosis,
although there were many uncertainties regarding the children's fluoride exposure
(potential confounders included drinking tea high in fluoride, differences in nutritional
status, and one person obtaining water from sources with vastly different water fluoride
concentrations). Also, the fact that there is fluorosis at all levels, even the <0.4 mg/L,
raises a question about the source of the fluoride. However, Table 3 does demonstrate
that the portion of the population with the severe fluorosis increases as the
concentration of the water increases so there is some dose response.
Dental fluorosis
271
January, 2008
-------
Wondwossen, F., A.N. Astrem, K. Bjorvatn, and A. Bardsen. 2004. The relationship between dental
caries and dental fluorosis in areas with moderate- and high fluoride drinking water in
Ethiopia. Community Dent. Oral Epidemiol. 32:337-344.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL
POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL
METHODS:
STUDY DESIGN
PARAMETERS
MONITORED:
Dental fluorosis and dental caries
Prevalence survey
Ethiopia/Rift Valley. 306 Ethiopian children 12-15 yrs old (152 girls, mean age 13.5 yrs, and
154 boys, mean age 13.1 yrs), from three neighboring villages in the Rift Valley; the 3
villages are all within the Wonji Shoa Sugar Estate and are stated by Wondwossen et al
(2004) to be of approximate same size and socioeconomic condition. Informed consent was
obtained from participating children, their parents and local authorities. The authors point out
that boys and girls were unevenly represented in the two study areas (but equally represented
in the total sample).
None
Beginning in 1982, and lasting for 12-15 years; the dental examinations took place in 1997.
The children were grouped into two exposure categories based on fluoride drinking water
levels measured in water wells in the three villages, as shown in the Table 1 taken from
Wondwossen et al. (2004).
Tal'li1 I \ vi': 11141' fkiti-'tlr ioiM.viitr.it:i>:i i:nt;/'^ mi\«s-
urrt.1 an tht' df:nking i\nti't ul Jit1 iilsnit uio- .uid iii^li-
lliirif jdf .ijt'.l
Vutii \5<>cU').Hi-ilj)'r!i-fku>i;ih'
,Kst»-~me:il aiiM j:tv
HV-F«: C. i-l 1 •JO-l 1.1
I.ISI-KW t: .!-l f> H.I-I ] i
i>r, ft ">-i 9 inn-1 1 1
i 't»r a •;-._: 2 u;j;-i lit
The children in the study population resided in villages that relied on local ground water
sources for their drinking water. Exposure to fluoride through other routes such as air or diet,
were not assessed in this study; therefore, total fluoride intake was not reported.
Not reported. Fluoride levels in ground water sources were assessed between 1982 and 1997.
Dental fluorosis and caries incidence were evaluated in 306 children 12-15 yrs old, from three
neighboring villages in the Rift Valley, Ethiopia. The children were grouped into two
exposure categories based on well-water fluoride concentrations measured in 1982-1997
(average values 0.4-1.4 mg/L to 0.3-2.2 mg/1 for the moderate exposure group and 0.89-14. 1
mg/L to 10.0-14.0 mg/L for the high exposure group). Dental fluorosis was scored using the
Thylstrup-Fejerskov (TF) Index, and dental caries was measured with DMFS indices and
DMFT scores. Statistical analysis consisted of bivariate analyses using cross-tabulation, chi
square statistics, independent sample t test, one-way ANOVA, Spearman's correlation
coefficient (ys), and the Kruskal-Wallis test.
Scoring for fluorosis was based on the Thylstrup-Fejerskov (TF) Index (see Section 2) and
assessed on the vestibular, occlusal and lingual surfaces; scoring for dental caries was based
272
January, 2008
-------
STATISTICAL
METHODS:
RESULTS:
Dental fluorosis
Other effects
on the DMFS indices and DMFT scores as described by WHO Oral Health Survey system
(see Section 2). DMFT scores for each individual were established by summing highest
DMFS scores across all permanent teeth.
Data were analyzed using the Statistical Package for the Social Sciences (SPSS v. 10.0).
Bivariate analyses were performed using cross-tabulation, chi square statistics, independent
sample t test, one-way ANOVA, Spearman's correlation coefficient (ys), and the Kruskal-
Wallis test. Dental caries prevalence was regressed on TF scores and area of residence,
controlling for confounding factors by the use of multiple logistics regression analyses.
Ninety-five percent confidence interval (95% CI) was given for the odds ratio.
Prevalence of dental fluorosis (TF>1) was 91.8% in the moderate fluoride areas and 100% in
the high fluoride area. When compared with 12-year olds with TF scores 0-4, odds ratios for
DMFT >1 was 3.0 (95% CI = 1.6-5.7). For children 13-15 years olds with TF scores >5,
odds ratio for DMFT >1 was 2.0 (95% CI = 1.2-3.2). No statistically significant interaction
effect on caries was identified for the term TF scores and place of residence, indicating that
the strength of association between dental caries and dental fluorosis did not systematically
vary between moderate and high-fluoride areas.
rn -i
•£.4;?,- H D Mode rate-
is 40- | fluorid* an?*
2 35 - | H High-
•g 30 - i* fluoride area
o 1:0 - *
*S" [I n H ^
I In.i a.ir!.rU B.
TFU TF1 TF2 TF3 TF4 TFr- TFi, TF7
Median TF-score for the individual
ftj I Hie I'Wti'iK.ny Jistrjb.itui:! u: ,: <,K".
H't't'.i' :n niitdiT itt1- >':iJ "lU'^Juflun-ak1 jrf, ,HIL| flu,ir*j^t^ € jnt - uuK , 'Ml I
Mthji'tiit.Mlu irtjf in- l,i }^*\ ."> (r'4'-i -*! (-ir' 'i
HH'.II flu >tkiu jn-.i i,,, : ;i ,'i-v..,) ,^(H>,,|
I itil I.; i'i»ii %-J (4S"i lt;r 1 = ] "i
•OcU li H r.l-i ' 2-i ^ -W
-j ir'i •- ,AJ± 2. A,
1M?h'i 6(3.9) 1.47 ±2.04
273
January, 2008
-------
Mr in PM"-'I 'i>iti" jth! -I inJ<-(>1 thulium in vk-iiti! Miiiit'i-i-- it Jiitrmil Ji,H;ti,i4t, iiit.il! [Ottil- in
mi) hvvh liu.'EiJr irt'd-
' ilu
i tiu•(- * 1 4"
+>> t- 2 "5
16
29
67
112
M. t '~I
114
87
89
306
Degree of fluorosis in relationship to tooth type as shown in Figure 1 taken directly from
Wondwossen et al, 2004
fa}
as
|i
i s
W
I
2 ^~
I*
a.
1TF1-2
DTP 3-4
QTF5-?
2nd moiar 1st molar pfsmoiaps canfnss
90-i
80-
70-
60-
50-
40-
30-
20-
10-
QTF3-4
DTFB-7
canines incisors
Tooih ty$m
I (*' 2 lhs* |H'ftitilsnr Js-Eilinlinn
M rtd,i[i-.,.-nl<. w..itilnn; t» PMI'l
!^^ - jcil 1 -, «(M~u I'xHhhp in
Ihr nsiili't ill' 111 HiJ hll',1] lluuriJ*
(bl ,m>i
STUDY AUTHORS'
CONCLUSIONS:
The prevalence of dental fluorosis (TF > 1) was 91.8% (moderate fluoride area) and 100%
(high-fluoride area).
A positive relationship between caries and fluorosis was observed across tooth types in both
the moderate- and high-fluoride areas areas. The percentage of children with DMFT > 1 was
highest in groups with TF score > 5 in the 2?d molar, followed by the 1st molar. The 2nd
molar was the tooth most severely affected by dental fluorosis in the age group.
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE
NOT FOUND IN NRC
(2006)
For definitions and descriptions of scales and indices, please see Section 2 and List of
Acronyms.
PROFILER'S
REMARKS
DMO
11/14/06
and
12/15/2006
Because of the wide range of fluoride levels identified in the two exposure groups, the data
cannot be used to accurately assess dose response relationships. However, assuming that a
TFI of 5 and above is equivalent to severe fluorosis, then the proportion of the population
showing severe fluorosis can be estimated to have been about 63% in the high fluoride area,
and about 13% in the low fluoride area.
A previous dental fluorosis study (Olsson 1979) identified the presence of some defluoridated
water supply systems in Wonji-area villages; it is unclear if these defluoridated water supplies
have been incorporated into the present study.
274
January, 2008
-------
1
PROFILER'S ESTIM.
NOEL/NOAEL
PROFILER'S ESTIM.
LOEL/ LOAEL
POTENTIAL
SUITABILITY FOR
DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
Dental fluorosis observed at all fluoride concentrations examined; thus, estimating a NOAEL
for fluorosis is not possible from these data.
TBD
Not suitable (J, Poor (_), Medium (X), Strong (J
Dental fluorosis and caries
275
January, 2008
-------
Wondwossen, F, A.N. Astrem, K. Bjorvatn, and A. Bardsen. 2006. Sociodemographic and
behavioural correlates of severe dental fluorosis. Internat. J. Paed. Dent. 16:95-103.
ENDPOINT STUDIED:
TYPE OF STUDY:
POPULATION STUDIED:
CONTROL POPULATION:
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE ASSESSMENT:
ANALYTICAL METHODS:
PARAMETERS
MONITORED:
STATISTICAL METHODS:
RESULTS:
Caries
Dental fluorosis
Dental fluorosis
Prevalence survey; continuation of a study reported by Wondwossen et al 2004.
Ethiopia/Rift Valley: 306 Ethiopian children 12-15 yrs old (152 girls, mean age 13.5 yrs,
and 154 boys, mean age 13.1 yrs), from three neighboring villages in the Rift Valley; the 3
villages are all within the Wonji Shoa Sugar Estate (WSSE) and, according to Wondwossen
et al (2004), are of approximate the same size and socioeconomic condition. The dental
examinations took place in 1997. Informed consent was obtained from participating
children, their parents and local authorities.
None
Lifetime, up until the time of the dental examinations in 1997.
The study population was separated into two groups; villages A and M were in the moderate
fluoride area with fluoride levels ranging from 0.3 to 2.2 mg/L; and village K was considered
to be in a high fluoride area with fluoride levels of 10 and 14 mg/L. The authors point out
that boys and girls were unevenly represented in the two study groups, but equally
represented in the total sample.
Factors evaluated included fluoride concentration in drinking water derived from wells; use
of fluoride supplements and dentifrices; diet, including tea and fish consumption; breast-
feeding; and use of clay cooking pots. Total fluoride intake, however, was not estimated.
Information on fluoride levels in drinking water (obtained from drilled wells) was provided
by the medical service of the WSSE; however, analytical methods were not reported.
Fluorosis was assessed on the buccal, occlusal and lingual surfaces according to the
Thylstrup-Fejerskov (TF) Index (see Section 2 for description).
The SPSS (Statistical package for Social Sciences) software for PC, Version 10-0,
computer program was used in the analysis. Bivariate analyses were performed using
cross-tabulation, chi-square statistics and one-way analysis of variance. The Mann-
Whitney U -test and Kruskal-Wallis test were used when comparing severity according to
selected independent variables. Multiple logistic regression analysis was used to estimate
the risk for severe dental fluorosis, calculating odds ratio (ORs) and 95% confidence
interval (CI). Collinearity between the independent variables was checked by means of the
variance inflation factor (VTF), using xt as the dependent variable in regression on the other
independent variables. All analyses are based on the total number of participating
mother/child pairs (n= 233).
Not evaluated
Fluorosis, as measured by the TFI scores, was significantly greater (pO.OOl) in the high
fluoride area, as indicated in Table 2 taken from Wondwossen et al (2006). Higher levels
of fluorosis were also associated with: 1) shorter periods of breast feeding, tea and fish
consumption; and storage of drinking water in non-clay containers.
276
January, 2008
-------
STUDY AUTHORS'
CONCLUSIONS:
1 it*k- 2 Medi *r Thvl cmj,~ K~J,-I 4,,,1- ln,Ks i11E< .nics ittr 1 j iH rhn,j qf sink-i in .ill Kcth, cjrl- mipie, ! SeUli an j 1 tic f mpte 1 !f-cth
M," If in TH M,-n. tin-*! iiturtiL fhiid LfLuiasici
'*a nhle f-lmnlcr '"II !-eth Fash f-nsplini! auh 1 'Ee uiiptiuu icc-th
\rci "1 s*° Mien* B
mid. rate fluundc H2 2 tl i 1 II > "i IJil'l ". (>i 2lMli>4>ii
hiOi flifiMt. xl Siii4ii SHI" 45i4.n i2r' Sm4J. s^i--
A',!>- IVCTF\I
12 "2 2 ll(l H 4> i"* ! 5 . II, .l"i" 2 11 il « 4tir-
1 ,-1% 141 4n,j«,--iJ, 4n, s 4*1 4111211*11!
Stm-l<°r
nuk 121 ' II s M *i Mi* In, It ^»*i* 4 li 4n, 1 " 1 1 -" "ill,
Pr," < ikcdinii lini-nth !
M |H ,\'i 4"",2li ^>MS' 4** i "1 l-'j*" ^'M^M s;ihvv
- U 144 2 5 il n 4>li 2n ^ u, 4'M >!• il ^ 4 1i
Tea i ' 'nHi uplit^ii
dnik, f-iJjiU III' 45i|<; "Ii-i" 4H. ( 5 Id- • 4 ". (2 1' 5 Hi* -
•tfU.ni n. vi .IrinL li; 2^.111 4»i 2u, il j%, ^MilS 4'">i
^ivh f ii iHiipti'jii
nit il lilUi HI < II i 1 li 4s! i 2H i 1 II 4"H T; (I 1 ^ In
K.II hll t ml h.iic Mi 41i'7 Sh* 45(2,< < Sj 5m4n'>li'
Ju ri 1 i.jt 22 2 ' il n '< < 2 u .Ha .m ~» il n 42i
Chilli 1)t ' tt. 1
L!J- |.n 48 1-5(1-0; 4-3) ITM'B, 4 (M 2>M|.| 511,
inelilli |Ia n jj5j 3.0 (1-S; 5-8)* 111 fill 4'i' 41,1, Xil SHi*
*P«:0-i'r, *' IMHI
Among the 233 children, the prevalence of severe dental fluorosis (TFI > 5) was 24.1% and 75.9% in
the moderate- and high-fluoride areas, respectively (see Table 3, copied directly from Wondwossen
et al 2006):
hMt * F ult li 1>K! LI i vi i m h,r j a hnlit tin ->!M i I i^d MI Ibc tO'Tlnn I t th^ y|id JcntiU, n Th* d r*n,l"nl miHI-
i>*4lmuij i iln nitlnn Th 1 H»r-I liil'^iTHi a * jnjllini lun Tf ! . 11 _'
Ntni"!% 11 p r nf !4^ li 113 TI I ^ n" —
llni r n Unl in il k \lln,K i ' 1 U, riti , titi.l n n,t-i il ] > i <]
"i i it i i kn
m v) in flui n 1 |t_ 21 1 (1 '1
hlJinifcrid 81 26-1 10-3-66-0 ~5V (r,m
is i \ u i
12 92 2^ ", t2H'
P 1 • 141 2-6 1-1-6-0 74" i>'i
* "nd< i
mile 121. 0-9 04-2-1 ullXi4Hi
t, mil 112 - - T»2 i'I .
1'r 5 ti-tjm^ Ini ntfe):
il |H 89 4-7 2-4-9-4 f.i ^ nt'l
U 144 - - lii 7i2'o
T j Lt'OHiniph >n
JimkvJjil 110 3-6 1-5-8-& u'Mi.^i
HJ an n.- u Junk at ill 12.1 - - .>"4 i24i
1 1 h > t ri^ujnjin >n
m il lilH 181 0.8 0-2-2- S "UOi'.hi
m^il hllct inJ 1 line 38 1-2 0-3-4-0 24lll"l
li.n.t -jl 22 - 1 1 i4i
ilnri^^ , t 'at'"?
h- pi-!. 48, - - l~ 7 l!4>
riBtallK |,[ i fi, 185 2-9 1-1-7-9 SO!in'"i
1 ilh i f» ^ujutnin
h 1 1 fi l, t i,il.i 167 - - 7"" i5,,i
uth-r liil hh ur,-r 66 l-l 0-4-3-0 2wli2^i
M nlhh in * in iPiii
|i,H il.i>) 46 0-3 0-1-1-2 If,'" iMl
H 1- Jin inn. Immi 150 2-4 O-S-7-1 747i5"i
nl ibiah 37 - - K'i MI
According to bivariate as well as multivariate analyses, a number of sociodemographic and
behavioural factors were related to severe fluorosis. The odds for having severe fluorosis varied
according to the fluoride concentration of the drinking water, age, consumption of tea, length of
breastfeeding and method of storing water. The adjusted odds ratios ranged from 2.6 to 26. 1 .
Breastfeeding for > 1 8 months and the use of clay pots for storing drinking water was associated with
significantly lower TFI scores. Bivariate analyses indicated that being male and consuming fish
might be associated with higher TFI scores.
In order to avoid dental fluorosis, low-fluoride drinking water should be provided in the relevant
villages. A prolonged period of breastfeeding, the use of clay pots for storing water, and possibly a
reduced intake of tea and whole fish in infants might also help to avoid severe fluorosis in children
growing up in traditionally fluoride-endemic areas. The authors consider this report to be the first
211
January, 2008
-------
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE NOT
FOUND IN NRC (2006)
PROFILER'S Initials/date
REMARKS DMO/1/12/07
PROFILER'S ESTEM.
NOEL/NOAEL
PROFILER'S ESTEM.
LOEL/ LOAEL
POTENTIAL SUITABILITY
FOR DOSE-RESPONSE
MODELING:
CRITICAL EFFECT(S):
documentation of a relationship between clinical dental fluorosis and use of clay pots for water
storage (e.g., fluoride binding capacity of various clays) Bjorvatn et al (2003).
Bjortn, K., Reimann, C., 0stvold, S.H., Tekle-Haimanot, R., Melaku, Z. Siewers, U. 2003.
High fluoride drinking water. A health problem in the Ethiopian Rift Valley. 1 Assessment of
lateritic soils as defloridating agents. Oral Health and Preventive Dentistry 1:141-148.
The study showed that fluoride levels of 10 and 14. 1 mg/L in drinking water were associated
with the occurrence of severe fluorosis (TFI score > 5) in 75% of the exposed population.
Can not be determined
Severe fluorosis (median TFI of 5) occurred in the high fluoride area, associated with
fluoride levels in drinking water of 10 and 14. 1 mg/L.
Not suitable ( ), Poor (X), Medium ( ), Strong ( )
Only two exposure groups were evaluated, and the one for the moderate fluoride area
included a range of fluoride drinking water levels (0.3 to 2.2 mg/L) too wide to be useful in a
dose-response analysis.
Dental fluorosis
278
January, 2008
-------
Yoder, K.M., Mabelya, L., Robison, V.A., Dunipace, A.J., Brizendine, E.J., Stookey, G.K. 1998. Severe dental fluorosis
in a Tanzanian population consuming water with negligible fluoride concentration. Community Dent. Oral Epidemiol.
Dec; 26(6):382-93.
ENDPOINT STUDIED:
TYPE OF STUDY:
CONTROL
POPULATION:
(Site 1)
POPULATION STUDIED:
(Site 2)
POPULATION STUDIED:
(Site 3)
EXPOSURE PERIOD:
EXPOSURE GROUPS:
EXPOSURE
ASSESSMENT:
ANALYTICAL
METHODS:
STUDY DESIGN
Dental fluorosis and caries
Cohort
84 school children of heterogeneous tribal ethnicity (>12 tribes represented), ages 9-19 years old, from the
Chanika school in Tanzania near sea level (100 m), where the fluoride level in the drinking water is
negligible (0.046±0.047 mg/1).
100 school children of heterogeneous tribal ethnicity (>12 tribes represented), ages 9-19 years old, from the
Rundugai school in the plains of Tanzania at 840 m, where the fluoride level in the drinking water is
5.72±4.71 mg/1.
100 school children of homogenous tribal ethnicity (Chagga tribe), ages 9-19 years old, from the Kibosho
school located at 1463 m on Mount Kilimanjaro in Tanzania where the fluoride level in the drinking water is
negligible (0.18±0.32 mg/1).
Subjects were lifelong residents, so exposure period ranged from 9 to 19 years.
284 subjects were randomly selected from class rosters from three sites: Chanika School (Site 1, 100 m
altitude, negligible water fluoride), Rundugai School (Site 1, 840 m altitude, high water fluoride level), and
Kibosho School (Site 3, 1463 m altitude, negligible water fluoride). The residents of all three sites were
primarily farmers and not nomadic and were required to be lifelong residents of the study site from which
they were recruited, to be in good health, and to consent to participation.
Subjects were examined for dental fluorosis and caries. Subjects were interviewed about their food habits,
environmental characteristics, and use of a fluoride-containing food tenderizer known locally as magadi.
Parents were not questioned.
Meal (n=280), urine (n=280), water (n=42) and magadi (n=139) samples supplied by participants were
analyzed for fluoride content at the Oral Health Research Institute of Indiana University School of Dentistry.
Magadi samples from Sites 1 and 3 (n=2 from each site, randomly selected from 98 total samples) were
analyzed for complete element composition at the Indiana State Department of Health, Environmental
Laboratory. Water, urine, and magadi were directly analyzed for fluoride using a fluoride ion-specific
electrode (Orion Research, Boston, MA). Foods were analyzed by a modification of the diffusion method of
Taves (1968). Urine samples also were analyzed for creatinine concentration to enable determination of the
urinary fluoride to urinary creatinine ratio as a means of correcting for collecting spot urine specimens rather
than 24-hour samples; creatinine was determined at the Indiana University Hospital Endocrinology
Laboratory.
283 school children, ages 9 to 19, were examined in three locations of varying altitude in Tanzania during
1996. Three sites, representing high and low altitude with negligible fluoride and a mid-altitude site with
high water fluoride were selected from records of the Ministry of Health and Social Welfare of the United
Republic of Tanzania.
The children were questioned by a Tanzanian interviewer regarding length of residence at that site; their
consumption of tea, fish, and milk; the family's use of magadi; the family's cooking location (inside or
outside the home) and type of fuel used; frequency of use of insecticide on crops; and their younger siblings'
ingestion of tea and food cooked with magadi. Subjects were asked to list what they ate the previous day in
order to survey the types of foods consumed in that village. Information regarding nutritional status, on the
district level, was obtained from records of the Regional Maternal and Child Health Office, relating height
and weight to WHO standards.
Subjects were instructed to bring a sample of their evening meal (as a mixture, excluding drink) and a
sample of their first morning urine void in two separate closable 8-oz containers. They also brought a sample
279
January, 2008
-------
(-tablespoonful) of magadi used in their home. Meal, urine, water and magadi samples were analyzed for
fluoride content. Urine samples also were analyzed for creatinine concentration. Four randomly selected
magadi samples from Sites 1 and 3 were analyzed for complete element composition.
Examinations: Subjects were examined for dental fluorosis using the Thylstrup and Fejerskov Index (TFI)
and the Tooth Surface Index of Fluorosis (TSIF). Dental caries was recorded as decayed, missing and filled
permanent surfaces (DMFS) by Radike's criteria. One examiner quantified TFI scores and DMFS rates at all
sites; another examiner examined TSIF scores at Site 1 where minimal fluorosis existed; and a third
examiner examined for TSIF scores at Sites 2 and 3 where severe fluorosis existed. TSIF examinations were
conducted without drying teeth, and then the next examiner used gauze to dry the teeth for the TFI
examinations. Intraoral mirrors and explorers and battery-operated head lamps were used. Aseptic
procedures were followed. No radiographs were taken during the surveys. Intra-rater reliability for TSIF and
TFI fluorosis scoring was measured by random and blind re-examination of 10% of the subjects. Kappa
statistic was 0.74, 0.89, and 0.60 for the three examiners regarding fluorosis scores, within acceptable limits.
PARAMETERS
MONITORED:
Dental fluorosis was measured using the Thylstrup and Fejerskov Index (TFI) and the Tooth Surface Index
of Fluorosis (TSIF). Dental caries was recorded as decayed, missing and filled permanent surfaces (DMFS)
by Radike's criteria. A qualifier (S) was added to both fluorosis indices to indicate an anterior tooth
appearing to be artificially abraded for the purpose of removing pitting/discoloration due to fluorosis. A
qualifier (A) was added to TSIF scores to indicate an occlusal surface which had more abrasion than would
be expected in a child of that age.
STATISTICAL
METHODS:
Statistical methods were not reported.
RESULTS:
Dental fluorosis
All tables were copied directly from Yoder et al. (1998). Table 1 was split for size considerations and
summarizes specimen fluoride values, urine creatinine values, and mean maximum fluorosis scores (water
fluoride concretion and fluorosis scores presented here). Water fluoride concentrations at Site 2 varied
dramatically (range: 1.26 to 12.36 mg/L, mean: 5.72 mg/L). Sites 1 and 3 had negligible water fluoride
concentrations (0.05 and 0.18 mg/L, respectively). Mean TFI (TSIF) scores were 0.01 (0.01), 4.44 (3.14),
and 4.39 (3.59) for Sites 1, 2, and 3, respectively. As expected, subjects from Site 2, who consumed high
levels of fluoride in the drinking water, exhibited severe fluorosis. Yet, although both Sites 1 and 3 had
negligible water fluoride, Site 1 children had minimal fluorosis while virtually all subjects at Site 3
experienced severe fluorosis.
Table 1
««•• 'Ml
• ) *
* U
"t y ,* tiftiJuy.*'
,V*>B^J* *»4f' TTi
'•\^f,lf ft •• iW
'V*i
.-if.
t u f {"",4
»»£•»!?
*13»i
. > ;4
f V
< • : v.
Table 2 reports percentage distribution of TSIF fluorosis scores. At Site 1, less than 1% of subjects had
TSIF scores >1. Distribution was relatively similar at Sites 2 and 3. 85% of subjects with high TSIF scores
(6-7) had >4 molars with occlusal surfaces which were excessively abraded by wear; in contrast, subjects
with low TSIF scores (>3) did not have similar abrasion
280
January, 2008
-------
Caries
Fluoride levels
Other factors: dietary
taMf 3. PfrewUff dMrtbutimt t4 fiu-orosts scare* hi *R MI*$W»* >cc*h Surface lndc< •« ilufrrews (IMF)
., , ln-,mr»uu«»«y 1uw*A<> w<«r*<%l
V>ri^>i*t
«it» ij{sufi«es f 1 2 3 4 S f» ,'
f rfr 1 *'lKt 0*0 PO? KM PO? (MU 0,30 (Hi)
SM feSSS IMS 1573 444 i'ft-W WW II *» lift 14*1
*»lc 1 «?*fe MU 14.89 <•'1
Distribution was relatively similar at Sites 1 and 3 . Approximately half of the subjects with high TFI scores
(7, 39%; 8, 51%; 9, 60%) artificially abraded their anterior teeth in an attempt to remove pitting and
discoloration.
liJ4» t fmwrMUft* djsrnbwuwm >• * ?Sw>*«K*4 Ksres Icr »1I frrth "HKvlurap FejwskiM lml*< <4s fTF!)
, (MMibutuin c>t ihionKti wtwi I**)
Kumiwr - , -^ ,.___,
Fwtr i5l t«:A n S 2 3 1 1 »» ,'81
R-lp 1 JC»7 *4SW OKI 019 fl
511*1 1«I7 fl»> 241 726 R.4* W^7 ' 14# 1M.? 1 X? 5 €3 ITU
Sit* 3 J«* !47 ",23 S ~* •4i4 t 5"*36 t* »1 luj^ j>14 »-tf
i J"
Caries experience was low in all three sites; only 43 of the 284 children had DMFS >0. Mean DMFS for all
sites combined was 0.52±1.52. At Sites 1, 2, and 3, meanDMFS was 1.39 (±2.45), 0.15 (±0.73), and 0.19
(±0.6 1), respectively. Sixty percent of caries was detected in the least fluorosed teeth (TFI 0-1) and the next
highest percentage of caries (16%) was detected in the most severely fluorosed teeth (TFI 8-9). 71
permanent teeth had 116 decayed surfaces, which were predominately in occlusal surfaces (54%).
PROFILER'S NOTE: A table summarizing caries data was not provided in the article.
Table 1 was split for size considerations and summarizes specimen fluoride values, urine creatinine values,
and mean maximum fluorosis scores. Mean urinary fluoride concentrations were 0.52, 4.43, and 1.43 mg/L
for subjects at Sites 1, 2 and 3, respectively. Children from Site 3 had unexpectedly high mean urinary
fluoride (1.43 mg/L) compared with fluoride in the water they consumed (0. 18 mg/L), but unexpectedly low
urinary fluoride compared with their mean fluorosis scores (TFI, 4.39; TSIF, 3.59). Eight (of 100) subjects
at Site 2 had urinary fluoride >10 mg/L, which is the reference value described as toxic by SmithKline
Beecham (1997-98), and an additional 5 subjects' urinary fluoride levels were 8-10 mg/L.
Magadi fluoride concentrations were highly variable (range 189 to 83,21 1 mg/L); mean concentrations were
16010, 41 13, and 5037 mg/L at Sites 1, 2, and 3, respectively. Fluoride concentration in meals was widely
variable, ranging from 0.03 to 22.04 mg/L. Mean fluoride in meal samples was 0.49, 2.47, and 2.14 mg/L at
Sites 1, 2, and 3, respectively; meal fluoride was considered high at Sites 2 and 3.
There were statistically significant linear relationships between food fluoride and urine fluoride (r=0.3 1) as
well as between urine fluoride and magadi fluoride (p <0.000 1) (Pearson correlation coefficient, combining
all sites). Thus, urine collected was a good indicator of the fluoride content in the food consumed the
previous night.
*V,«* ; lan^sfi <•*' %^r, '^JCT *& ,^ ^j I ;»!^ "•^jt.mv-ft-i Bt^'»ft y$ **^«?n
IM?**!
*WVi %»«(i».li Wmi dvtil* C.-WW.W *.-«< ,M4'.fliiw
'<'> flwt.U frtf," (UMiJt!r> li fta>!i iftrt, 't • -Vf t. t.iy ;<»! ft r i -,>»*. ijiM 'irM'
Vf 1 i'jinn Hi |t . p ».,«i ifii" -,.« r K|
iMi'v\(i.' W"fM'J*W> »'f>,'4t-,l »" ».i'i 1 HH* H«-.j".;»' t, Slv* i»%i,|
"i*.m» v / Hi, :»>Ui!» *'11 n •»: r»u>. »•.!» ••!,>* * H"
4'Mui)» W' m .r",4* tfl j.'ip,! f 4jjv t«:-l*A t"i%*jW t-l"*J t-(\4*
•j,,-«!«4,t« "u" i- ,-;n« t",'«;j: i-*?",1 *»*j»t ( MSV
•-. .j^ ,.jj; ,,, ^
s.v»>tf jtc^w. -«%»!)"» t.-i.'s'vi n«i -i: ^-Z'tc1* «^^^?n4 * w?i fin;*
Consumption of milk, tea and fish varied from site to site, but statistical differences were not reported. Milk
281
January, 2008
-------
factors, magadi use,
nutritional status
STUDY AUTHORS'
CONCLUSIONS:
DEFINITIONS AND
REFERENCES CITED IN
PROFILE THAT ARE
NOT FOUND IN NRC
(2006)
PROFILER'S Initials/date
REMARKS SJG/
10/23/07
consumption appeared most variable with a low of 1 .4 half-cups/ week at Site 1 (negligible fluorosis) to a
high of 5 half-cups/ week at Site 2. 74% of subjects at Site 1 reported that they never drank milk. Tea
consumption ranged from 6.23 (Site 3) to 8.66 (Site 1) cups/week. Fish consumption ranged from 4.24 (Site
3) to 6.42 (Site 2) times/ month.
Magadi was used by 77, 96, and 99% of families in Sites 1, 2, and 3, respectively. The families cooked with
magadi a mean of 1.50, 2.24, and 3.04 times/week. When asked if babies and toddlers were fed food cooked
with magadi, 9.5, 89, and 100% respondents at Sites 1, 2, and 3, respectively, said "yes". Element analysis
revealed that some components of magadi were aluminium (Al), iron (Fe), magnesium (Mg), manganese
(Mn), molybdenum (Mo), strontium (Sr), and titanium (Ti), with much higher concentrations of these
elements in samples from Site 3 as compared with Site 1; Al 7.6 times, Fe 54 times, Mg 5 times, Mn 8
times, Mo 2 times, Sr 4.7 times, and Ti 7.2 times higher in Site 3 as in Site 1. Bioavailability was not
determined. Al and Mg are two elements that have potential as risk factors for enamel disturbances (human
effects have not yet been proven).
Approximately half of the children from the regions of Sites 2 (53%) and 3 (44%) were considered
malnourished by WHO standards between 1981 and 1990, which were the years when their permanent teeth
would have been forming. Data for Site Iwere not available.
An analysis of covariance model showed that all three communities differed significantly in mean fluorosis
scores (pO.OOOl). Controlling for urinary fluoride concentration and urinary fluoride: creatinine ratio,
location appeared to significantly affect fluorosis severity. Urinary fluoride: creatinine ratio had a stronger
correlation than urinary fluoride concentration with mean maximum fluorosis scores (r=0.43 vs r=0.25).
Fluorosis at Site 3 (high altitude) was more severe than would be expected from the low water and normal
urinary fluoride values. Altitude, the elements (including fluoride) contained in magadi, and other nutritional
factors may contribute to the severity of fluorosis observed. Although statistical analysis implicated lifelong
residence at that location (1436 m altitude) as a risk factor in the severe fluorosis which was observed at Site
3, it is reasonable to expect that other factors may contribute to the high prevalence of fluorosis. The
differences observed in the populations which could be potential risk factors for developing severe fluorosis
included the following:
o Frequency and patterns of use of magadi (fluoride-containing cooking additive)
o Elements, besides fluoride, in magadi (which were found in higher concentrations in magadi from
Site 3 vs. Site 1)
o Altitude of residence since birth
o Nutritional factors (malnutrition, tea and insufficient milk consumption)
o Genetic factors
Taves, D.R. (1968). Separation of fluoride by rapid diffusion using hexaamethyedisiloxane. Talanta 15:
969-974.
SmithKline Beecham Clinical Laboratories (1997-98). Directory of services reference guide.
Collegeville, Pennsylvania.
The study was well-conducted and had adequate study design. However, the study was not designed for
development of a dose response to fluoride as the emphasis was on monitoring dental fluorosis in children
and on attempting to explain other factors besides fluoride in the drinking water that may contribute to
fluorosis.
Not surprisingly, children from Site 1 (Chanika School, low altitude, negligible water fluoride) had very
little fluorosis (TFI and TSIF scores= 0.01), but the most caries (DMFS score = 1.39). Urinary fluoride
levels (0.52 mg/L) were consistent with water (0.05 mg/L) and meal fluoride levels (0.49 mg/L). The mean
magadi fluoride level was high (16010 mg/L), but magadi use was lowest in this group (77% of families, 1.5
times/week) and fewer babies and toddlers were fed food cooked with magadi (9.5% answered 'yes')
according to interviews.
Children from Site 2 (Rundugai School, mid-altitude, high water fluoride) had moderate fluorosis (TFI =
4.44; TSIF=3.14) and low caries experience (DMFS=0.15). Urinary fluoride levels (4.43 mg/L) were
consistent with water (5.72 mg/L) and meal fluoride levels (2.47 mg/L).
282
January, 2008
-------
Children from Site 3 (Kibosho School, high altitude, negligible water fluoride) had surprisingly high
fluorosis scores (TFI=4.39; TSIF=3.59) and urinary fluoride concentrations (1.43 mg/L) relative to water
fluoride levels (0.18 mg/L); however, fluoride from meals contributes to urinary fluoride levels, so urinary
fluoride levels actually were not surprising. Several factors that were common to the subjects at Site 3 have
been theoretically associated with severe dental fluorosis: residence at high altitude; relatively high
magnesium; excessive fluoride in food due to magadi use; and malnutrition. Caries experience also was low
(DMFS=0.19).
Limitations of the study:
o High altitude alters urinary excretion of fluoride, so urine may not be the best body fluid for
measuring fluoride body burden; however, blood sampling was not feasible in this study. Urine
samples from toddlers (during amelogenesis) could have provided additional information, but this
was beyond the scope of the current study.
o Element analysis was conducted on only 2 magadi samples each from Sites 1 and 3, so results are
suggestive but not conclusive.
o Statistics methods were not described.
o Malnutrition data were not available for Site 1.
PROFILER'S ESTEM.
NOEL/NOAEL
Study design was not suitable for development of a NOAEL for fluorosis or caries.
PROFILER'S ESTEM.
LOEL/ LOAEL
Study design was not suitable for development of a LOAEL for fluorosis or caries.
POTENTIAL
SUITABILITY FOR
DOSE-RESPONSE
MODELING:
Not suitable (X), Poor (), Medium (), Strong ()
While the study was well-conducted, the study design was not conducive to provide data for a dose-
response. The study indicated that factors besides drinking water fluoride levels contribute to fluorosis
prevalence and severity (e.g., altitude, food additive, nutrition). The study did not address any issues of
plaque or gingivitis.
CRITICAL EFFECT(S):
Severity of dental fluorosis; caries experience
283
January, 2008
------- |