Australian Research Centre for Population Oral Health,School of Dentistry, The University of Adelaide, Australia5005; *corresponding author, loc.do@adelaide.edu.au
J Dent Res86(8):723-728, 2007
ABSTRACT
This study aimed to evaluate the risk-benefitbalance of several fluoride exposures. Fluorideexposure history of randomly selected childrenwas collected for calculation of exposure tofluoridated water, toothpaste, and other fluoridesources. We evaluated the risk-benefit balance offluoride exposure by comparing dental fluorosison maxillary central incisors, recorded at the timeof the study with the use of the Thylstrup andFejerskov Index, and deciduous caries experience,recorded at age six years, of the same group ofSouth Australian children who were from 8 to 13years old in 2002-03. Population Attributable Riskfor fluorosis and Population Prevented Fractionfor caries were estimated. Fluorosis prevalencewas found to be 11.3%; caries prevalence, 32.3%;mean dmfs, 1.57 (SD 3.3). Exposure to fluoridatedwater was positively associated with fluorosis, butwas negatively associated with caries. Using1000-ppm-F toothpaste (compared with 400- to550-ppm-F toothpaste) and eating/lickingtoothpaste were associated with higher risk offluorosis without additional benefit in cariesprotection. Evaluation of the risk-benefit balanceof fluoride exposure provides evidence to assist inthe formulation of appropriate guidelines forfluoride use.
KEY WORDS: fluoride, fluorosis, caries, risk-benefit balance, children.
Received November 24, 2005; Last revision March 14,2007; Accepted April 15, 2007
A supplemental appendix to this article is publishedelectronically only at http://www.dentalresearch.org.
Risk-Benefit Balance in the Use ofFluoride among Young Children
INTRODUCTION
T
he use of fluoride for promoting oral health has always involved abalance between the protective benefit against caries and the risk ofdeveloping fluorosis. Monitoring fluoride exposure in childhood continuesto be important in preserving the effectiveness of fluorides in cariesprevention, while limiting the risk of fluorosis.
Fluoride exposure during the first few years of life has a potential bi-directional association with oral health. On the benefit side, early fluorideexposure protects newly erupting deciduous teeth, creating a \"biologicallyfriendly\" oral environment. It is suggested that this deciduous cariesprotection is an important outcome of early fluoride exposure per se, and isa predictor of later permanent caries experience (Li and Wang, 2002; Skeieet al.fluorosis (Fejerskov , 2006). On the risk side, early exposure to fluoride is a risk factor foret al., 1994).
The current generation of children is exposed to numerous fluoridesources, each of which has an unknown balance of benefit and risk.Identifying and maintaining an effective balance of the caries-protectivebenefit and the fluorosis risk are crucial to the confidence of the dentalprofession and the population in the use of fluoride. Epidemiologicalevaluation of the risk and benefit balance is often based on measurement ofthe hazardous and protective impact of an exposure (in this case, fluorideexposure) (Spassof, 1999). At the population level, the hazardous impact[where Relative Risk (RR) > 1] can be measured as Population AttributableRisk (PAR), which defines the proportion of cases (children with fluorosis)attributed to the exposure in the population. The protective impact (RR < 1)can be measured as Population Prevented Fraction (PPF), which defines theproportion of cases (children with caries) prevented by the exposure in thepopulation. These impact indicators are of interest in epidemiologicalevaluation, because they measure the absolute impact of an exposure on adisease or condition in a population (Rose, 1994).
This study aimed to evaluate the balance of benefit and risk of severalfluoride exposures among South Australian children.
METHODS
The study was nested in a large-scale population-based study, the Child OralHealth Study (COHS), conducted in 2002–04 among South Australian childrenattending the School Dental Service (SDS). The study design and data collectionprocess have been detailed elsewhere (Do and Spencer, in press).
Three sources of data were collected for each child: fluoride exposurehistory from the COHS questionnaire collected in 2002-03, caries experiencefrom SDS-archived clinical records, and fluorosis experience collected in 2003-04.
Caries experience data were extracted from SDS-archived clinical records(APPENDIX). For the purpose of this study, we used data recorded at the firstavailable dental visit after a child turned six years old. We used only caries
723
724
Table 1.Study Samples' Characteristics
Do & SpencerJ Dent Res 86(8) 2007
according to the TF Fluorosis Index,based on the photographs after thefieldwork completion, so that intra-BoysGirlsTotal
examiner reliability could becalculated. The absolute agreementTotal sample, n (row %)349 (51.6)328 (48.4)667 (100)
was above 80%, and the kappascores ranged from 0.74 to 0.79 forExposure to F in water, birth to age 3 yrsa(n = 627, col %)
each and both central incisors> 50% lifetime59 (24.2)56 (23.8)115 (24.0)
combined.> 0-50% lifetime116 (51.6)112 (49.7)228 (50.7)
We re-weighted the data to0% lifetime145 (24.2)139 (26.6)284 (25.3)
adjust for different sampling ratiosand age and sex distribution. TheAge when fluoridated toothpaste use commenced (n = 596, col %)
<18 mosweights were used in the analysis to136 (40.1)141 (47.7)277 (43.8)
produce representative estimates for19-30 mos98 (34.9)89 (31.8)187 (33.4)
the South Australian child31+ mos71 (25.0)61 (20.5)132 (22.8)
population.
Caries experience, recordedToothbrushing when F toothpaste use startedb
when a child was six years old, andUsed 400- to 550-ppm-F toothpaste (total n = 610); n (%)*214 (69.4)188 (59.3)402 (64.4)
fluorosis prevalence, recorded at theBrushed 2+ times/day (total n = 611); n (%)189 (41.9)192 (38.1)381 (35.6)
examination in 2003-04, wereSwallowed slurry after brushing (total n = 589); n (%)*183 (57.0)146 (47.1)329 (52.1)
compared between and amongUsed small amount of toothpaste (total n = 609); n (%)100 (32.6)92 (31.1)192 (31.8)
groups with different levels ofHaving an eating/licking toothpaste habit (total n = 609); n (%)*145 (44.5)172 (53.4)317 (48.9)
fluoride exposure in bivariateanalyses. The case definition ofPrevalence of fluorosis (TF Fluorosis Index 2+)c(total = 667); n (%)* 21 ( 8.3)36 (14.7)57 (11.3)
caries for multivariate logistic
dregression models was thePrevalence of deciduous caries, age 6 yrs(total = 603); n (%)142 (34.2)121 (30.4)263 (32.3)
prevalence of caries at age six years,
dand the case definition of fluorosisMean deciduous dmfs at age 6 yrs(total = 603); mean (SD)1.63 (3.3)1.50 (3.3)1.57 (3.3)
was having a TF Fluorosis Index
*Chi-square, sex comparison, p < 0.05.score of 2+ on one or both maxillaryaEstimate as percent of life period exposed to fluoride in water.central incisors. We used fluoridebPatterns of toothbrushing practice when a child commenced fluoridated toothpaste use.
exposures to generate logisticcFluorosis defined as having a TF Fluorosis Index score of 2+ on one or both maxillary central incisors.
dregression models for the fluorosisCaries experience on deciduous molars and canines recorded when a child was six yrs old.
prevalence at the time of the studyand the caries prevalence at age sixyears, with the significance level set
experience on deciduous molars and canines to calculate theat 0.05. All fluoride exposures in early childhood were included inprevalence of caries and dmfs score among the children when theythe models, because all can theoretically be a risk factor forwere six years of age.fluorosis and a protective factor for caries. Sex, birth cohorts,
A fluoride exposure history of each child was collectedparental education, and household income were also included inthrough a 12-page self-administered parental questionnaire (Dothe models. Estimates of those logistic regression models wereand Spencer, in press). Percent lifetime exposure to fluoridatedused to estimate Population Attributable Risk (PAR) for fluorosiswater for the birth-to-age-three-years period was estimated andand Population Prevented Fraction (PPF) for caries, according to aused to categorize children into three groups: having 0% lifetime,method presented in the medical literature (Bruzzi et al., 1985).
<> 0 & 50% lifetime, and > 50% lifetime exposure. Age whenThis method has previously been used in fluorosis researchfluoridated toothpaste use began, and patterns of toothbrushing(Riordan, 1993; Pendrys, 2000). The number of potentiallypractice when toothpaste use began were collected and used in theaffected children was calculated as a function of PAR and PPF andanalysis (APPENDIX).the prevalence of fluorosis and caries.
COHS participants in South Australia, who were born inEthics approval was obtained from the University of Adelaide1989-94 inclusive (from 8 to 13 years old at the time of the study),Ethics Committee. Informed parental consent was received for useand whose questionnaire and clinical caries data were available,of caries data and for the children's participation in the fluorosiswere invited to be examined for fluorosis at their local SDS clinic.examination.One calibrated dentist (LGD) conducted all the examinations using
RESULTSthe Thylstrup and Fejerskov Index (Fejerskov et al., 1988) under
standard clinical conditions. Permanent teeth were cleaned andA total of 677 children participated in the study (Table 1). Adried with compressed air and scored for fluorosis. For thisquarter of the children had been exposed to fluoridated wateranalysis, a case of fluorosis was defined as having a Thylstrup &over half of their first three years, while another quarter had noFejerskov (TF) Fluorosis Index score of 2+ on one or bothexposure. Nearly half of the children commenced fluoridatedmaxillary central incisor.toothpaste use before 19 months, but some 23% commenced it
Children's front teeth were photographed with a clinicalafter 30 months of age. Two-thirds of the children reportedlydigital camera. The maxillary central incisors were scoredused a children's 400- to 550-ppm-F toothpaste. Half of the
J Dent Res 86(8) 2007Risk-Benefit Balance in the Use of Fluoride725
Table 2.Dental Caries and Fluorosis Experience of Children with Different Exposures to Fluoride (bivariatechildren reportedly had an
analysis)eating/licking toothpaste
habit.
Fluorosis ExperienceaCaries ExperienceA TF Fluorosis Index
Prevalencedmfs ScorebPrevalencecscore of 2+ was observed
%OR * (95% CI)Mean (SD)%OR * (95% CI)on the maxillary central
incisors of 57 children
Exposure to fluoridated water,(11.8%). Among those, 48
birth to age 3 yrs (n = 627)had the same fluorosis
†#0.9 (2.3)> 50% lifetime14.06.2 (2.2-17.9)25.50.6 (0.5-0.8)scores on contralateral
> 0-50% lifetime10.94.6 (1.6-13.7)#1.7 (3.4)30.10.7 (0.5-0.9)teeth. A TF Fluorosis Index
†0% lifetime3.412.2 (4.1)45.81score of 3 was the highest
severity score observed. A
Age when F toothpaste use started (n = 596)TF Fluorosis Index score of
<18 mos13.21.5 (1.0-2.6)† 1.0 (2.5)24.10.6 (0.5-0.8)1 (visible only after teeth
#19-30 mos12.21.4 (0.8-2.4)1.1 (2.6)29.60.6 (0.4-0.7)were dried) was observed in
†# 2.0 (4.0)30+ mos7.8147.01another 11.6%. A total of
603 (89%) children had had
Type of toothpaste usedd(n = 610)an archived clinical
Standard 1000-ppm-F toothpaste*16.21.8 (1.1-3.0)1.1 (2.8)29.80.9 (0.8-1.1)examination at age six
Children's 400- to 550-ppm-F toothpaste6.411.5 (3.1)32.71years [mean age, 6.3 yrs
(SD 0.4)]. The prevalence
Brushing frequencyd(n = 611)of deciduous caries was
† 1.1 (2.8)Twice a day or more11.01.0 (0.6-1.6)30.60.9 (0.7-1.2)32.3%, with a mean dmfs of
†1.5 (3.2)Once a day or less11.1133.511.6 (SD 3.3).
Exposure to fluoridated
After-brushing routined(n = 589)water from birth to age
Swallowed (with or without rinsing)13.01.1 (0.9-1.4)1.3 (2.9)28.80.9 (0.8-1.1)three years was
Spat out (with or without rinsing)10.311.8 (3.7)35.01significantly associated
with the prevalence of
Toothpaste amountd(n = 609)fluorosis and caries, in both
Medium or larger13.41.3 (0.8-2.2)1.4 (2.9)30.60.9 (0.7-1.2)bivariate (Table 2) and
Small amount10.611.8 (4.0)33.81multivariate analyses
(Table 3). The group with
Eating/licking toothpaste habitd(n = 609)0% exposure to fluoridated
Yes14.51.8 (1.1-3.0)1.4 (3.0)32.81.1 (0.9-1.3)water had a significantly
No8.711.4 (3.1)29.81lower prevalence of
fluorosis. However, this
aDefined as having one or both maxillary central incisor with a TF Fluorosis Index score 2+.group also had abMean deciduous dmfs at age 6 yrs (molars and canines only; SD in parentheses).
significantly higher†#One-way ANOVA, Tukey's post hoc test, indicating pairs that significantly differed; column comparison.prevalence and severity ofcPrevalence of deciduous caries at age 6 yrs (molars and canines only).
dcaries compared withPatterns of toothbrushing practice when a child commenced fluoridated toothpaste use.*OR, Crude Odds Ratio (95% CI); statistically significant if 95% CI did not include unity.groups with exposure to
fluoridated water. Thegroups exposed to
prevalence of fluorosis. However, using this type of toothpastefluoridated water from birth to age three years had significantly
was not associated with a significantly higher carieshigher odds of having fluorosis, but significantly lower odds of
prevalence, in both bivariate (Table 2) and multivariatehaving caries at age 6 years compared with the group with 0%
analyses (Table 3). Brushing 2+ times a day was not associatedexposure (Table 3).
with fluorosis, but it was associated with a lower mean dmfs,Age when fluoridated toothpaste use began was also
but not significantly lower caries prevalence at age 6 yearsassociated with both fluorosis and caries (Table 2).
(Tables 2, 3). In multivariate analysis, swallowing slurry afterCommencing toothbrushing after 30 months was not related to
brushing was associated with higher odds of having fluorosissignificantly lower fluorosis prevalence, but was related to
without a significant association with caries (Table 3). Havingsignificantly higher caries prevalence and severity, compared
an eating/licking toothpaste habit was associated withwith starting brushing before 30 months in either bivariate
significantly higher risk of having fluorosis, in both bivariate(Table 2) or multivariate analysis (Table 3). Commencing
(Table 2) and multivariate analyses (Table 3), without afluoride toothpaste use in the 19- to 30-month period was not
significant association with caries (Tables 2, 3).associated with higher caries experience, compared with earlier
The impact measures on fluorosis and caries of Southcommencement (Table 2).
Australian children are presented for fluoride exposures inThe use of a children's 400- to 550-ppm-F toothpaste when
Table 4. The Population Attributable Risk for fluorosis oftoothpaste use started was associated with a significantly lower
726Do & SpencerJ Dent Res 86(8) 2007
was not associated with a significant PPF againstcaries. Having an eating/licking toothpaste habit wasattributed to 44% of fluorosis cases (50 cases/1000children), without a significant association with cariesat age 6 years. Swallowing slurry after brushing wasattributed to 33% of fluorosis cases. Commencingtoothpaste use before 30 months of age was notassociated with a significantly higher risk of fluorosiscompared with later commencement. However, thatpattern was attributed to fewer caries cases at age 6years (29%, or 95 cases/1000 children).
Table 3.Logistic Regression Models for the Prevalence of Fluorosis and Caries byExposures to Fluoride in Childhood [Odds Ratio (95%CI)]
Logistic Regression ModelsFluorosisaCariesb
Explanatory Factors
Exposure to fluoridated water, birth to age 3 yrs
> 50% lifetime*5.6 (1.9–16.6)> 0-50% lifetime*4.7 (1.7–13.4)0% lifetime1Age when toothpaste use started<18 mos19-30 mos30+ mos
*0.4 (0.2-0.7)*0.5 (0.3-0.9)
1
ns1.3 (0.6-2.9)ns1.3 (0.6-3.0)
It is believed that this paper is the first studyreporting the impact measurements (PAR and PPF)of fluoride use for both fluorosis and caries in the
Toothpaste when toothpaste use started
same children. Findings of this study can help
1000-ppm fluoride toothpaste*2.9 (1.5–5.3)ns0.8 (0.5-1.4)
evaluate the impact of fluorides. The study used
400- to 550-ppm fluoride toothpaste11
population data specific for South Australianchildren. Therefore, generalizability of the study
Brushing frequency when toothpaste use started
findings is dependent on each population's fluoridens1.2 (0.7-2.2)ns1.0 (0.7-1.6)Twice/day or more
exposure profile, oral health status, and SES.
Once/day or less11
This study used complex data collectionprocedures. Retrospective and concurrent data were
Toothpaste amount when toothpaste use started
collected on inter-related aspects of oral health andns1.6 (0.9-2.9)ns1.1 (0.7-1.7)Medium amount or larger
their contributory factors. The comprehensive
Small amount11
questionnaire collected fluoride exposure history, tofacilitate estimation of fluoride exposures that could
After-brushing routine when toothpaste use started
be related to both fluorosis and caries prevalence.ns1.3 (0.8-2.2)Swallowed (with or without rinsing)*2.2 (1.2-4.2)
Recall error is an inherent limitation of
Spat out (with or without rinsing)11
retrospective research. The COHS questionnaire wascarefully designed to minimize such error. We tested
Eating/licking toothpaste habit when toothpaste use started
the reliability of recalled data by comparingns1.0 (0.6-1.6)Yes**3.4 (1.8–6.5)
estimates of this sample with the whole COHS
No11
sample in South Australia. The fluoride exposuredata of this study sample were similar to those
Parental education
reported by the parent COHS sample (unpublished),ns2.0 (1.0-4.2)Secondary school or less*1.8 (1.1-3.2)
and by a study in Western Australia (Riordan, 2002).ns0.9 (0.4-2.1)Trade/vocational*2.3 (1.3-4.2)
The examiner was blind to the questionnaire's
University11
data at the examination, but not blind to a child'scurrent residence. However, current residence wasaLogistic regression model for the prevalence of fluorosis, defined as having a TF
not used as a direct explanatory factor. AlthoughFluorosis Index score of 2+ on one or both maxillary central incisors (n = 554;
2pseudo R: 0.175).strongly related, current residence did not always
bLogistic regression model for the prevalence of deciduous caries at age 6 yrs (nreflect the exposure to fluoridated water during the
= 480; pseudo R2, 0.127).
birth-to-three-years age period used in this study.nsp > 0.05; *p < 0.05; **p < 0.001.
Caries data were collected by 31 uncalibratedOther factors in the two models: sex, birth cohort, fluoride supplements, infant
clinicians. However, the clinicians used uniformformula, and household income. Actual age in mos at the 6-year examination
was included in the model for caries prevalence.manuals to perform the examinations. Also, analyses
were based on the presence/absence of cavitatedcaries experience (either filled or not), which is
exposure to fluoridated water approximated 56%. This wasreliable (Rugg-Gunn et al.,1976).equivalent to 63 cases per1000 SA children. On the benefitCaries experience at age six years was used for theside, having exposure to fluoridated water prevented 34% ofanalysis, to coincide with the time when most children startcaries cases at age six. This was equivalent to 111 cases perenrolling in the SDS. Caries on deciduous molars and canines1000 children who were prevented by fluoridated water fromwas used because anterior teeth were exfoliating, andhaving caries at age 6 years.permanent caries experience was negligible at age six.
PAR estimates for fluorosis of several patterns of fluoridatedFluorosis experience on maxillary central incisors at ages 8 totoothpaste use ranged from 27 to 44% (Table 4). The use of13 yrs was used because all children would have those teethstandard 1000-ppm-F toothpaste was responsible for 27% ofpresent. Although limiting the observation to those teeth mightfluorosis cases, or 30 cases per1000 children, compared with theunderestimate the prevalence of fluorosis, maxillary centraluse of children's 400- to 550-ppm-F toothpaste. This exposureincisors are more appropriate, because they are developed
1
*0.4 (0.2-0.8)
*0.6 (0.4-1.0)
1
DISCUSSION
J Dent Res 86(8) 2007Risk-Benefit Balance in the Use of Fluoride727
Table 4.Population Attributable Risk (PAR), Population Prevented Fraction (PPF), and Potential Changes in Number of Cases
Fluorosis
Potential Changeb
63-30--3750
Caries
Potential Changed
11195-----
PARa
Exposed to fluoride in water, birth to age 3 yrsCommencing toothpaste use before 30 mos
Used 1000-ppm F compared with 400- to 550-ppm Ftoothpaste, when brushing started
Brushing twice/day or more, when brushing startedUsed a pea-sized or larger amount of toothpaste,when brushing started
Swallowing slurry after brushing, when brushing startedEating and/or licking toothpaste, when brushing started
PPFc
55.7 (37.9, 61.6)29.8 (-8.2, 43.8)26.8 (14.5, 33.4)6.7 (-19.0, 20.6)23.5 (-10.5, 42.1)32.6 (7.8, 45.6)44.3 (27.7, 53.1)
34.3 (5.7, 50.9)29.3 (7.9, 34.5)-5.5 (-27.0, 8.0)1.3 (-18.0, 13.5)5.3 (-31.2, 28.2)10.4 (-7.5, 21.5)-0.3 (-29.6, 19.1)
abcd
PAR and PPF were derived from logistic regression models. These estimates are not mutually exclusive; therefore, they do not add to 100.Fluorosis: Prevalence of fluorosis defined as having a TF Fluorosis Index score 2+ on one or both maxillary central incisors.Caries: Prevalence of deciduous caries (dmfs > 0) at age 6 yrs.
Population Attributable Risk (95% CI): Proportion of cases attributed to exposure.
Number of cases per1000 children having a TF Fluorosis Index score of 2+ attributed to the exposure, given the population prevalence of 11.3%.Population Prevented Fraction (95% CI): Proportion of cases prevented by exposure.
Number of cases per1000 children with deciduous caries at age 6 yrs prevented by the exposure, given the population prevalence of 32.3%.
(hence, at risk) during the first few years of life, when the issueof risk and benefit of fluoride exposure is most critical. Also,maxillary central incisors are the most observable teeth, giventhat the effect of mild fluorosis is esthetic. Future plans withthe sample will focus on collection and analysis of data onfluorosis and caries in all permanent teeth.
The risk-benefit balance of community water fluoridationhas long been established. Water fluoridation has been hailed asone of the 10 most successful public health achievements for itsbenefit in preventing caries (CDC, 2001). Such a benefit ofwater fluoridation has often been found associated with a risk offluorosis (Ismail et al., 1990; Szpunar and Burt, 1990; Riordanand Banks, 1991; Spencer et al., 1996; Mascarenhas, 2000).
This study reported unadjusted and adjusted estimates ofrisk and benefit of exposure to fluoridated water. There was asignificant association of exposure to fluoridated water withboth fluorosis and caries among the same children. Thissuggests that a balance between the risk and benefit ofexposure to fluoridated water already exists at the current levelof fluoridation. A significant deterioration of child oral healthwould occur if exposure to fluoridated water was reduced oreliminated in Australia. Potential impact of any such changeshould always be carefully evaluated.
Unlike water fluoridation, risk and benefit of toothpaste usecan be dependent on patterns of its use and the oral healthstatus of the children using it. The efficacy of fluoridated, ascompared with non-fluoridated, toothpaste in the prevention ofcaries has been clearly established (Marinho et al., 2003).However, fluoridated toothpaste is one of the main sources offluoride intake, thus contributing to a risk for fluorosis (Levy etal., 2001, 2003).
Since almost all the children in this study population usedfluoridated toothpaste, it was impossible to compare fluoridatedtoothpaste use with that of non-fluoridated toothpaste.However, components of fluoridated toothpaste use might stillbe modified to create a more favorable risk and benefit balance.Those components were: age when the use of toothpaste began,
type of toothpaste, after-brushing routine, and an eating/lickingtoothpaste habit. These findings suggested that there was anopportunity to refine fluoridated toothpaste to lower the risk offluorosis without significantly reducing its effectiveness incaries prevention. The measures might include encouragingcommencement of toothpaste use in the 19- to 30-month ageperiod, use of 400- to 550-ppm fluoridated toothpaste whentoothpaste use starts, encouraging spitting after brushing, and(especially) preventing an eating/licking toothpaste habit inyoung children. Appropriate recommendations for toothpasteuse and their timely dissemination to the public, especiallyparents of young children, would shift the balance toward moreprotection from caries without the unnecessary risk of fluorosis.Such recommendations are pertinent, given the significant riskof fluorosis attributable to those components of fluoridatedtoothpaste use.
To conclude, water fluoridation and the patterns offluoridated toothpaste use can have different risks and benefitsfor oral health. Appropriate guidelines that are based on theevaluation of the risk and benefit of each component of fluorideexposure can lead to a more beneficial outcome.
ACKNOWLEDGMENTS
The study was supported by the University of Adelaide, by aNational Health and Medical Research Council Project Grant,by an Australian Dental Research Foundation grant, and by theSouth Australian Dental Service. Professors John W. Stammand Andrew J. Rugg-Gunn and two anonymous Reviewers areacknowledged for their highly constructive comments.
REFERENCES
Bruzzi P, Green SB, Byar DP, Brinton LA, Schairer C (1985). Estimating
the population attributable risk for multiple risk factors using case-control data. Am J Epidemiol 122:904-814.
CDC (2001). Recommendations for using fluoride to prevent and control
dental caries in the United States. Centers for Disease Control andPrevention. MMWR Recomm Rep 50(RR-14):1-42.
728Do & SpencerJ Dent Res 86(8) 2007
Do LG, Spencer AJ (in press). Decline in the prevalence of dental fluorosis
among South Australian children. Community Dent Oral Epidemiol(accepted May, 2006).
Fejerskov O, Manji F, Baelum V (1988). Dental fluorosis: a handbook for
health workers. Copenhagen: Munksgaard.
Fejerskov O, Larsen MJ, Richards A, Baelum V (1994). Dental tissue
effects of fluoride. Adv Dent Res 8:15-31.
Ismail AI, Brodeur JM, Kavanagh M, Boisclair G, Tessier C, Picotte L
(1990). Prevalence of dental caries and dental fluorosis in students, 11-17 years of age, in fluoridated and non-fluoridated cities in Quebec.Caries Res 24:290-297.
Levy SM, Warren JJ, Davis CS, Kirchner HL, Kanellis MJ, Wefel JS
(2001). Patterns of fluoride intake from birth to 36 months. J PublicHealth Dent 61:70-77.
Levy SM, Warren JJ, Broffitt B (2003). Patterns of fluoride intake from 36
to 72 months of age. J Public Health Dent 63:211-220.
Li Y, Wang W (2002). Predicting caries in permanent teeth from caries in
primary teeth: an eight-year cohort study. J Dent Res 81:561-566.
Marinho VC, Higgins JP, Sheiham A, Logan S (2003). Fluoride toothpastes
for preventing dental caries in children and adolescents. CochraneDatabase Syst Rev (1):CD002278.
Mascarenhas AK (2000). Risk factors for dental fluorosis: a review of the
recent literature. Pediatr Dent 22:269-277.Pendrys DG (2000). Risk of enamel fluorosis in nonfluoridated and
optimally fluoridated populations: considerations for the dentalprofessional. J Am Dent Assoc 131:746-755.
Riordan PJ (1993). Dental fluorosis, dental caries and fluoride exposure
among 7-year-olds. Caries Res 27:71-77.
Riordan PJ (2002). Dental fluorosis decline after changes to supplement and
toothpaste regimens. Community Dent Oral Epidemiol 30:233-240.Riordan PJ, Banks JA (1991). Dental fluorosis and fluoride exposure in
Western Australia. J Dent Res 70:1022-1028.
Rose G (1994). The strategy in preventive medicine. Oxford: Oxford
University Press.
Rugg-Gunn AJ, Downer MC, Ashley FP (1976). Reliability of caries data in
three clinical trials. Community Dent Oral Epidemiol 4:15-18.
Skeie MS, Raadal M, Strand GV, Espelid I (2006). The relationship between
caries in the primary dentition at 5 years of age and permanent dentitionat 10 years of age—a longitudinal study. Int J Paediatr Dent 16:152-160.
Spassof RA (1999). Epidemiologic methods for health policy. New York:
Oxford University Press, Inc.
Spencer AJ, Slade GD, Davies M (1996). Water fluoridation in Australia.
Community Dent Health 13(Suppl 2):27-37.
Szpunar SM, Burt BA (1990). Fluoride exposure in Michigan
schoolchildren. J Public Health Dent 50:18-23.
因篇幅问题不能全部显示，请点此查看更多更全内容