Original Paper

Caries Res 2014;48:306–311 Received: May 23, 2013

Accepted after revision: September 10, 2013
DOI: 10.1159/000355613
Published online: February 14, 2014

Method for the Analysis of Total Fluoride

in Fluoride-Releasing Dental Varnishes
C.M. Carey S.S. Coleman
University of Colorado, School of Dental Medicine, Aurora, Colo., USA

Key Words Fluoride-releasing varnishes (F-varnish) are used to

Fluoride · Fluoride measurement · Fluoride varnishes help prevent dental caries [ADA, 2006; Lawrence et al.,
2008; Weintraub et al., 2006] and to treat dentin hyper-
sensitivity [Beltran-Aguilar et al., 2000; CDC, 2001]. The
Abstract number of F-varnish products in the US marketplace has
Today’s fluoride-releasing varnishes (F-varnish) contain a increased from 4 in 2003 [Hazelrigg et al., 2003] to more
wide variety of ingredients which present analytical chal- than 35 in 2013 [Dental Product Shopper, 2013]. Fluo-
lenges for measuring their total fluoride content. This study ride-releasing dental varnishes typically contain 5% so-
reports improved methods to measure fluoride content in dium fluoride (NaF, 2.26% F, 22,600 μg F/g) or 1% diflu-
F-varnishes. Six different commercially available F-varnishes orosilane (0.1% F, 1,000 μg F/g) as their active ingredient
that contain difluorosilane (0.1% F) or NaF (2.26% F) alone or suspended in a variety of resin/rosin mixtures that yields
in combination with calcium-phosphates were analyzed. In a range of handling properties. Other forms of fluoride
a vial, 1–3 drops (0.05–0.15 g) of varnish product was dis- such as sodium monofluorophosphate (MFP), amine flu-
pensed, dissolved in chloroform, equilibrated in TISAB and oride (NH3-F), potassium fluoride (KF), calcium fluoride
analyzed via fluoride ion-selective electrode. The average (CaF2), stannous fluoride (SnF2) or a combination of
weight percentage of fluoride for all F-varnishes containing these compounds could be envisioned within the varnish
NaF ranged from 2.03 to 2.24% F, which is within 90% of the product. There are also a growing number of F-varnish
declared label concentration of 2.26% F. Analysis of the diflu- products that contain adjunct ingredients such as xylitol,
orosilane-containing product required an additional hydro- flavorings, and a variety of calcium-phosphate (CaPO4)
lysis step. ANOVA found no significant difference between salts that are included to enhance fluoride release and up-
the 5% NaF varnishes at p < 0.05. This method for fluoride take into enamel. These CaPO4 additives have the possi-
analysis yields reliable and reproducible results and can be bility to interfere with the separation of fluoride for the
used for a wide variety of F-varnishes. The standard uncer- analysis of the product. National and international stan-
tainty for this method is ±4%. This method may become the dards organizations (ADA/ANSI and ISO) on dental
basis for national and international standards that ensure products are preparing new standards that will include
the F-varnish products used in clinical practice have the fluo- requirements on fluoride content for the total fluoride in
ride content declared in the product literature. the product. There is a need for the development of a
© 2014 S. Karger AG, Basel simple, sensitive, and rapid analytical method that can be
132.248.9.8 - 8/4/2014 5:44:18 AM
Dir.General de Bibliotecas, UNAM

© 2014 S. Karger AG, Basel Clifton Carey, PhD

0008–6568/14/0484–0306$39.50/0 University of Colorado, School of Dental Medicine
12800 E. 19th Avenue, MS 8310
Downloaded by:

E-Mail [email protected]
Aurora, CO 80045 (USA)
www.karger.com/cre
E-Mail clifton.carey @ ucdenver.edu
utilized as the standard method of analysis for total fluo- Methods
ride in F-releasing dental varnishes. This method may be-
Sample Preparation
come the basis for national and international standards The method was evaluated by measurement of total fluoride in
that ensure the F-varnish products used in clinical prac- six types of commercially available fluoride-releasing dental var-
tice have the fluoride content declared in the product lit- nishes that contained difluorosilane or sodium fluoride either
erature. alone or in combination with calcium phosphate salts. Three of
Shen and Autio-Gold [2002] described the measure- the products analyzed contained CaPO4 salts as tricalcium phos-
phate, amorphous calcium phosphate, or a mixture of calcium sul-
ment of total fluoride in three fluoride varnishes and re- fate and potassium phosphate. Other potentially confounding in-
ported a wide range of concentrations. This wide range gredients included protein, wax, xylitol, flavorings, colorings,
was either due to a lack of product uniformity or inherent thickeners and solvents such as ethanol, ethyl acetate, hexane, and
standard uncertainties related to the analytical methods glycerin.
used, or a combination of both. Since that time a large The six varnish types employed in this investigation included
two products that contained sodium fluoride alone (NaF-A and
number of new F-varnish products have been introduced NaF-B), that had very different carrier matrices. Three contained
to the marketplace that are designed to release fluoride sodium fluoride in combination with calcium phosphate: NaF
faster, reside on the tooth longer, or to have mobility with amorphous calcium phosphate forming salts (NaF/ACP),
around the oral cavity. Because F-varnishes have such a NaF with tricalcium phosphate (NaF/TCP), and NaF with casein
wide variety of compositions the development of a single phosphopeptide-amorphous calcium phosphate complex (NaF/
CCP-ACP). The sixth varnish contained 0.1% difluorosilane
method that can accurately determine the total fluoride (F2H2Si) as its fluoride source. Samples were taken from the indi-
content in commercially available F-varnishes is difficult. vidual varnish types within a single lot or reference number and
This is why we feel it is important to provide methods that within the label expiration assigned by the manufacturer.
can be used to assess the fluoride content for this wide To achieve product homogeneity, we mixed the varnish in the
variety of products. packaging prior to dispensing aliquots of varnish product. Sam-
ples were mixed a second time using either a stainless-steel spat-
We developed and tested a simplified sample prepara- ula or the applicator brush provided before dispensing the varnish
tion that can be used to analyze the fluoride content of product to the digestion vial. For the product packaged in squeeze
complex samples. The basic approach is to dissolve the tubes, after kneading the tube for a few minutes to mix the prod-
nonaqueous portion of the sample (resins and rosins) uct as best as possible, approximately 1 inch of product was dis-
which does not contain active fluoride into chloroform, a pensed and discarded from every new tube to ensure product ho-
mogeneity; samples were obtained from the remaining varnish
hydrophobic organic phase. The sample-chloroform product.
mixture is then mixed with an aqueous buffer to dissolve
the fluoride complexes into an aqueous phase. The buffer Sample Digestion
serves several functions including chelating cations to re- After testing, we determined that the required ingredients for
lease fluoride ions from ion pairs, and adjusting the ionic sample preparation should be added in the following order: First,
on a freshly tared balance, we added 0.05–0.15 g (1–3 drops) of the
strength and pH to create the optimum conditions for varnish product to a chloroform-resistant scintillation vial with
fluoride ion-selective electrode analysis. The buffer we 25 ml or greater capacity. We found that adding the sample di-
have found that meets our needs is marketed as Total Ion- rectly to the vial allows for a decreased loss via evaporation and
ic Strength Adjusting Buffer II (TISAB®-II, Sigma-Al- resulted in an easier and more accurate record of varnish mass. The
drich, St. Louis, Mo., USA). TISAB-II contains high con- F2H2Si varnish samples required this immediate attention after
dispensing when determining the F-varnish mass because the sol-
centrations of cation chelators which facilitate hydrolysis vent of this product evaporates quickly. We placed approximately
of fluoride-ion pairs by releasing fluoride ions from any 4 ml (±0.5 ml) of chloroform to each sample vial and swirled the
cation-fluoride complexes that may be in the product. An contents to dissolve the varnish. After the varnish product was dis-
equivalent buffer solution can be made following the rec- solved completely, we added 10.0 ml (±0.1 ml) TISAB to dissolve
ipe described in ASTM International [2010]. This sample and hydrolyze the fluoride complexes in the products. We ob-
served that the addition of TISAB polymerizes the resin in some
preparation sets our method apart from other methods in of the varnish products resulting in the formation of small beads.
the literature. Our goal was to develop a reliable method We determined that at this point in the dissolution phase, all con-
to measure total fluoride concentration [F] in a wide va- tents should be vigorously shaken by hand until all of these hydro-
riety of NaF and fluorosilane-based F-varnishes. We ana- phobic beads of resin are disrupted. Following this mixing, the
lyzed six types of F-varnishes which, according to pack- vials were stirred at 550 rpm overnight at room temperature. The
following day, the stir plate was turned off and the samples were
age labels, contain 5% sodium fluoride (2.26% fluoride) left to sit for at least 2 h to allow the chloroform and TISAB to
with the exception of the fluorosilane product, which separate. The top aqueous solution layer was then ready for fluo-
contains 1% difluorosilane (0.1% fluoride). ride measurement.
132.248.9.8 - 8/4/2014 5:44:18 AM
Dir.General de Bibliotecas, UNAM

Method for Measurement of Fluoride in Caries Res 2014;48:306–311 307

F-Varnishes DOI: 10.1159/000355613
Downloaded by:
 We found that after the separation phase of the NaF/ACP and Table 1. Fluoride concentration in varnish products
NaF-A samples exhibited three layers, the chloroform layer, a
white colored layer in the aqueous phase, and a clear aqueous Varnish [F], % NaF, % n3
phase. Prior to measurement, the two aqueous layers were lightly
mixed ensuring supernatant homogeneity. After testing we deter- average SD average SD
mined that if the cloudy layer forms within the chloroform layer,
as occurs with the NaF/CCP-ACP samples, only the supernatant NaF-A 2.03 0.04 4.48 0.10 05
layer should be analyzed. NaF/ACP 2.23 0.04 4.94 0.10 09
For one product we found that the rosin did not fully dissolve NaF/TCP1 2.20 0.29 4.87 0.64 09
in the chloroform, leaving small clumps of residue at the interface F2H2Si2 0.076 0.004 no NaF no NaF 10
of the chloroform and aqueous layers. The addition of an addi- F2H2Si-KOH 0.100 0.002 no NaF no NaF 06
tional 4.0 ml of chloroform and an additional mixing step of vigor- NaF-B 2.03 0.07 4.48 0.16 08
ously shaking prior to stirring overnight did not dissolve the NaF/CCP-ACP 2.24 0.03 4.95 0.07 09
clumps. Follow-up analysis of these clumps found that they did not
contain significant amounts of fluoride. 1
 The large standard deviation (SD) may reflect product inho-
For the F2H2Si samples we found that the addition of 1.0 ml mogeneity.
1 mol/l potassium hydroxide (KOH) solution was required as an 2 These F H Si samples were not hydrolyzed with base, the low-
2 2
initial preparatory step to hydrolyze the fluorosilane prior to the er measured fluoride concentration reflects incomplete hydrolysis
addition of 10 ml TISAB for the dissolution of fluoride ion. of the difluorosilane.
3
  n values are different due to combining data for products
Sample Analysis measured on several occasions and measurement of products
A fluoride ion-selective electrode was calibrated at room tem- which were acquired later in the study.
perature following the manufacturer’s instructions over the range
of 1 × 10–5 to 1 × 10–2 mol/l NaF. The calibration plot, log[F mol/l]
versus potential (mV), was linear and had a slope that fell within
the range of –57 to –59 mV per decade [F–] mol/l at room tem-
perature.
For the sample analysis, a blank solution containing 10.0 ml concentration. During the separation phase, a third layer
deionized water plus 9.0 ml TISAB solution was prepared for each of resin residues forms within the chloroform layer of
sample to be analyzed. The blank solution to be used for the F2H2Si many F-varnishes. For example, as a result of the varnish
samples was made with 9.9 ml water plus 9.1 ml TISAB solution to chemistry that allows for a protective coating of the cal-
maintain the same dilution as the other sample types. The aqueous cium within the varnish, a thick white gelatinous layer
phase of all samples was gently mixed prior to sample analysis to
ensure supernatant homogeneity; 1.0 ml of the mixed varnish sam- within the chloroform layer was observed in the NaF/
ple supernatant was added to the blank solution and analyzed via TCP varnish. Figure 1 shows the separated layers for the
the calibrated standard fluoride ion-selective electrode. The fluo- products.
ride concentration of the diluted sample was calculated from the Paired two-sample for means t test found that the total
standard curve and that value multiplied by the dilution factor of fluoride concentration values for three varnishes (NaF-A,
100 and divided by the mass (g) of the varnish in the sample to
calculate the fluoride concentration of the product in moles F/gram
 
NaF-B, and F2H2Si) were significantly lower than the stat-
product. The fluoride concentration (mol/g) multiplied by 19,000 ed label values (p < 0.05). The significant differences for
yields the fluoride concentration in micrograms F per gram of the NaF-A and NaF-B products are a result of small vari-
product (ppm F); this quantity can be multiplied by 2.21 to calcu- ations in the repeated sample measures. The current draft
late the concentration of sodium fluoride in micrograms NaF per of international standard ISO/DIS17730 ‘Dentistry – Flu-
gram of product (ppm NaF).
A paired two-sample for means t test was used to determine if oride Varnishes’ requires that the total fluoride content
there were significant differences between the measured total fluo- in F-varnishes do not deviate by more than 20% from the
ride concentration and the concentration value listed on the label. stated amount on the package. Even though the F-var-
ANOVA was used to determine if there were significant differ- nishes had less fluoride than stated on the labels the mea-
ences in the mean total fluoride concentration between compara- sured values for the NaF-A and NaF-B products were
ble 5% NaF-containing products.
within 10% of the label value and thus well within the ac-
ceptable range. The F2H2Si F-varnish was not treated with
KOH prior to fluoride extraction; the fluoride concentra-
Results tion of the samples that were treated with KOH was not
significantly different from the product label concentra-
The mean percent of total fluoride within commer- tion (p > 0.05). ANOVA found no significant difference
cially available F-varnishes is given in table 1. All products in the fluoride concentrations between the 5% NaF var-
contained at least 90% of the label amount of fluoride nishes (p < 0.05).
132.248.9.8 - 8/4/2014 5:44:18 AM
Dir.General de Bibliotecas, UNAM

308 Caries Res 2014;48:306–311 Carey/Coleman

DOI: 10.1159/000355613
Downloaded by:
 Color version available online
Fig. 1. Separation layers of the varnish products. A = Aqueous layer; Ia = insoluble layer at the aqueous interface;
Ic = insoluble layer at the chloroform interface; Is = insoluble layer in the chloroform layer; C = chloroform layer.

Discussion In the development of this method we investigated

several options for the protocol. When determining the
The method is shown to be acceptable for the analysis order of adding ingredients for the sample digestion we
of a wide variety of F-varnishes. The use of multifunc- thought that we could combine the chloroform and
tional TISAB to chelate fluoride-binding cations, as well TISAB together in a vial to stir the mixture rapidly to cre-
as providing ionic strength and pH buffering to extract ate a vortex, and then add the varnish product into the
the fluoride from the products simplifies the analysis. vortex. We hoped to avoid the polymerization reaction
TISAB contains high concentrations of 1,2-cyclohexyl- that occurs when some F-varnish products contact water.
enediaminetetraacetic acid and acetic acid that are cation Unfortunately, this approach made it difficult to deter-
chelators which would otherwise form cation-fluoride mine the exact mass of F-varnish dispensed into the vor-
complexes. The chelation of the cations releases fluoride tex, and it did not completely avoid the polymerization
ions from any fluoride ion pairs that may be in the prod- upon contact with the aqueous phase. Others have evalu-
uct. Additionally, the TISAB provides the ionic strength ated adding the F-varnish product directly to a mixture
and pH buffering needed for the ion-selective electrode of TISAB and chloroform or to the chloroform only and
analysis of total fluoride concentrations. The use of TISAB they report that adding the product to the chloroform be-
for the sample digestion and for fluoride ion-selective fore adding TISAB yielded better results than adding the
electrode analysis simplifies the method and reduces F-product to the TISAB-chloroform mixture [Flanigan
methodological errors. et al., 2013]. We found that the determination of the mass
132.248.9.8 - 8/4/2014 5:44:18 AM
Dir.General de Bibliotecas, UNAM

Method for Measurement of Fluoride in Caries Res 2014;48:306–311 309

F-Varnishes DOI: 10.1159/000355613
Downloaded by:
of F-varnish added to the chloroform was difficult be- mass of the sample to be analyzed can introduce as much
cause the chloroform had significant evaporation during as 2% error with a variation of ±0.002 g mass of F-varnish.
the time we weighed the sample. This introduced a large The method indicates determination of the F-varnish
error in the analysis, which leads to an overestimate of the product mass to ±0.0001 g; however, some samples have
fluoride concentration. a very volatile carrier making it difficult to measure the
We found that stirring the sample for an extended pe- amount of F-varnish dispensed with high accuracy. The
riod of time (overnight or at least 8 h) greatly reduced other potential source of error in the method is the mea-
variability. It is probable that any fluoride binding to cat- surement of fluoride concentration via fluoride ion-selec-
ions or entrapment within the nonaqueous polymers tive electrode. The fluoride ion-selective electrode ana-
takes a longer time to free the fluoride. We allow the sam- lytical system will drift as much as 1 mV over the course
ples to separate for several hours before taking aliquots of of several hours. At the typical concentration of the di-
the aqueous layer for analysis. As seen in figure 1 the sep- luted sample a ±1 mV drift will cause a variation of 900
aration of the samples often yields three layers. We ana- μg/g for samples of 22,600 μg/g fluoride concentration for
lyzed the compositions of the three layers and found that approximately ±4% standard uncertainty. Each of these
for the NaF/ACP product the white layer between the sources of error (or uncertainty) can be minimized
chloroform and TISAB contained fluoride presumably as through careful sample handling during the preparation
calcium-fluoride complexes which had settled. This layer for the sample digestion, careful sample pipetting and in-
was easily mixed into the TISAB layer with gentle mixing cluding measurement of fluoride standards interspersed
of the sample, the fluoride concentrations measured for between sample measurements. Due to these sources of
these samples were not significantly different from the error and fluoride ion-selective electrode variations over
concentration declared on the product label. The white time it is reasonable to assume that the standard uncer-
layer that appeared to be between the chloroform and tainty of this method is ±4%, which is equal to the stan-
TISAB for the NaF/CCP-ACP product was actually in the dard deviation of the repeated measures for the samples.
chloroform and seems to be proteinaceous in nature. This We suggest the inclusion of an internal standard at 5%
layer did not contain fluoride. Finally, we did not find any NaF (50,000 μg/g NaF), which is treated in exactly the
significant amounts of fluoride in the chloroform layer same manner as the F-varnish samples as a form of qual-
even if there appeared to be insoluble F-varnish compo- ity control for the method.
nents in that layer. This possibility was reported by Shen This method for the determination of total fluoride
and Autio-Gold [2002] when they first dissolved the concentration in F-varnishes yields reliable and repro-
product (package and all) into chloroform and then ducible results and can be used for a wide variety of F-
mixed the solution with water. varnishes. The estimate for standard uncertainty as de-
An additional hydrolysis step was needed for the prod- rived from an evaluation of uncertainties and the average
uct that contains difluorosilane. We found that the addi- standard deviation of the analyses is ±4%. This method
tion of 1.0 ml of 1 mol/l KOH prior to stirring overnight may become the basis for national and international stan-
in the preparatory stages of this method allowed the fluo- dards that ensures the F-varnish products used in clinical
ride ions to become available for measurement via fluo- practice have the fluoride content declared in the product
ride ion-selective electrode. It is anticipated that as new literature.
F-varnish products are introduced into the marketplace
there will be ingredients such as difluorosilane that will
require additional steps in the sample preparation phase Acknowledgments
of the analysis. This method is designed such that addi- This study was supported by the University of Colorado, School
tional steps could be added to the protocol without great- of Dental Medicine and a grant from the US National Institutes of
ly changing the overall analytical method of separating Health R01-DE021391. The funders had no role in study design,
the nonaqueous portions of the product from the water- data collection and analysis, decision to publish, or preparation of
soluble portion and using the strong chelating buffer to the manuscript.
hydrolyze any fluoride ion pairs for the analysis of the
samples.
Disclosure Statement
An analysis of errors from evaluation of each step of
the procedure shows that there are several steps where The authors declare that they have no conflict of interest asso-
measurement is critical. The first step of measuring the ciated with the present study.
132.248.9.8 - 8/4/2014 5:44:18 AM
Dir.General de Bibliotecas, UNAM

310 Caries Res 2014;48:306–311 Carey/Coleman

DOI: 10.1159/000355613
Downloaded by:
 References
ADA Report of the Council on Scientific Affairs: CDC: Recommendations for using fluoride to Lawrence HP, Binquis D, Douglas J, McKeown L,
Evidence-Based Clinical Recommendations: prevent and control dental caries in the Switzer B, Figueiredo R, Laporte A: A 2-year
Professionally Applied Topical Fluoride. May United States. MMWR Recomm Rep 2001; community trial of fluoride varnish for the
2006. 50(RR-14):1–42. prevention of early childhood caries. Commu-
ASTM International: Standard Test Methods for Dental Product Shopper: Varnish Products. nity Dent Oral Epidemiol 2008;36:503–516.
Fluoride Ion in Water. 2010, D1179-10, Sec- http://www.dentalproductshopper.com/ Shen C, Autio-Gold J: Assessing fluoride concentra-
tion 18.1. category/varnish/products, May 8, 2013. tion uniformity and fluoride release from three
Beltran-Aguilar ED, Goldstein JW, Lockwood Flanigan P, Vang F, Fitch J, Bigham W: Influence of varnishes. J Am Dent Assoc 2002;133:176–182.
SA: Fluoride varnishes – a review of their clin- dissolution method and calcium on varnish to- Weintraub JA, Ramos-Gomez F, June B: Fluoride
ical use, cariostatic mechanism, efficacy and tal fluoride. J Dent Res 2013;92(special issue A): varnish efficacy in preventing early childhood
safety. J Am Dent Assoc 2000;131:589–596. abstract 130 (www.dentalresearch.org). caries. J Dent Res 2006;85:172–176.
Hazelrigg CO, Dean JA, Fontana M: Fluoride var-
nish concentration gradient and its effect on
enamel demineralization. Pediatr Dent 2003;
25:119–126.

132.248.9.8 - 8/4/2014 5:44:18 AM

Dir.General de Bibliotecas, UNAM

Method for Measurement of Fluoride in Caries Res 2014;48:306–311 311

F-Varnishes DOI: 10.1159/000355613
Downloaded by: