Content uploaded by Jan Mulder
All content in this area was uploaded by Jan Mulder
Content may be subject to copyright.
Received April 27, 2008; Last revision April 8, 2009;
Accepted April 9, 2009
R. Yee1,4*, C. Holmgren1, J. Mulder1,
D. Lama2, D. Walker3, and
W. van Palenstein Helderman1
1Department of Global Oral Health, Radboud University
Nijmegen Medical Centre, College of Dental Sciences, PO
Box 9101, 6500 HB Nijmegen, The Netherlands; 2The
United Mission to Nepal Oral Health Programme, PO
Box 126, Kathmandu, Nepal; 3Population Oral Health,
Faculty of Dentistry, University of Sydney, 1 Mons Road,
Westmead, NSW 2145, Australia; and 4National University
Health System (NUHS), Faculty of Dentistry, Department
of Preventive Dentistry, Level 2, Dentistry Block, 5 Lower
Kent Ridge Road, Singapore 119074; *corresponding author,
J Dent Res 88(7):644-647, 2009
Arresting Caries Treatment (ACT) has been pro-
posed to manage untreated dental caries in chil-
dren. This prospective randomized clinical trial
investigated the caries-arresting effectiveness of a
single spot application of: (1) 38% silver diamine
fluoride (SDF) with tannic acid as a reducing
agent; (2) 38% SDF alone; (3) 12% SDF alone;
and (4) no SDF application in primary teeth of 976
Nepalese schoolchildren. The a priori null hypoth-
esis was that the different treatments have no
effect in arresting active cavitated caries. Only the
single application of 38% SDF with or without
tannic acid was effective in arresting caries after
6 months (4.5 and 4.2 mean number of arrested
surfaces; p < 0.001), after 1 year (4.1 and 3.4; p <
0.001), and after 2 years (2.2 and 2.1; p < 0.01).
Tannic acid conferred no additional benefit. ACT
with 38% SDF provides an alternative when restor-
ative treatment for primary teeth is not an option.
KEY WORDS: fluorides, silver diamine fluoride,
clinical trial, arresting caries treatment.
Efficacy of Silver Diamine Fluoride
for Arresting Caries Treatment
Untreated dental caries is a global pandemic (Edelstein, 2006). Due to lim-
ited financial resources, poor access to basic oral care, and the high cost
of restorative treatment, children of low-income nations have their general
health, social well-being, and education opportunities affected by untreated
dental caries (Baelum et al., 2007). Arresting Caries Treatment (ACT) has
been proposed to manage untreated dental caries in children of disadvantaged
communities (Bedi and Sardo-Infirri, 1999). Silver diamine fluoride (SDF),
Ag(NH3)2F, has been used to arrest caries since 1969 (Nishino et al., 1969;
Nishino and Yoshido, 1969; Yamaga and Yokomizo, 1969). Yearly applications
of 38% SDF (44,800 ppm F) to decayed primary anterior teeth of Chinese pre-
school children have been shown to be significantly more effective in arresting
caries and preventing new caries than three-monthly applications of sodium
fluoride varnish (22,600 ppm F) (Lo et al., 2001; Chu et al., 2002). In a Cuban
study, 38% SDF applied every 6 mos demonstrated clinical effectiveness in
arresting caries and preventing new caries in the dentition of 6- to 15-year-old
schoolchildren over a three-year period (Llodra et al., 2005).
Currently, the optimal frequency of application of SDF is unknown, but in
resource-limited situations, repeated applications of SDF are unlikely to be
either practical or affordable to local communities, even when applied by
trained primary-health-care workers.
Therefore, this current study compares the effectiveness of a single spot
application of two concentrations of SDF, 38% or 12%, in arresting caries,
with or without the use of tannic acid from tea as a reducing agent. The a
priori null hypothesis is that the different treatments have no effect in arrest-
ing active cavitated caries.
MATERIALS & METHODS
This prospective randomized clinical trial on a cohort of 976 kindergarten and
primary schoolchildren, 545 males (56%) and 431 females (44%), with ages
ranging from 3-9 yrs and a mean age of 5.2 (SD = 1.2) yrs at time of enroll-
ment, was conducted in Kathmandu, Nepal, a city with low-fluoride-content
drinking water (0.03 ppm). Fluoridated toothpaste is available from retailers,
but the children received neither professionally applied fluorides nor fluoride
supplements. Treatment was provided between May and August, 2005, and
the children were re-examined after 6 mos, 1 yr, and 2 yrs. The study was
approved by the Ethics Committee of the Nepal Health Research Council.
Written information explaining the study was sent to the parents. As well as
obtaining a written consent from the parents, verbal consent was obtained
from the children prior to commencement of the study. The children were at
liberty to withdraw from the study at any time during the study, and the same
J Dent Res 88(7) 2009 Efficacy of Silver Diamine Fluoride 645
freedom was afforded to the parents to withdraw their children
from the study.
The SPSS statistical program (version 10.0) was used for
random allocation of the children to one of the treatment groups/
Group 1: One application of 38% SDF for 2 min without a
Group 2: One application of 38% SDF for 2 min, with tea as
a reducing agent.
Group 3: One application of 12% SDF for 2 min, without a
Group 4: No treatment for carious teeth. This was the control
We used 38% SDF (44,800 ppm F) (Bee Brand Medical
Dental Company Ltd., Osaka, Japan) and 12% SDF (PROBEM -
Lab. Prod. Farmacêuticos e Odontológicos LTDA, São Paulo,
Brazil) to treat the teeth. We prepared tannic acid by slowly
boiling 2 teabags (Mechhi Tea; grown and manufactured in
Nepal) in 300 mL of water for 10 min.
Primary-health-care workers were trained to carry out the
treatment and collect information on the children’s oral health
behavior, quality of life, and socio-economic status. Treatment
assignment appeared on the children’s treatment plans, and
treatment was supervised by a dentist (RY). Prior to treatment,
the children brushed their teeth with a toothbrush without tooth-
paste and rinsed with water. The carious primary teeth were
isolated and kept dry with cotton rolls, and any debris was
removed with cotton pellets. No caries was removed. One drop
of SDF was placed in a plastic dappen dish, and a Vivabrush
#533664 type (Ivoclar Vivadent, Schaan, Liechtenstein) was
used to apply the SDF to carious surfaces of the primary teeth
for 2 min. If the protocol required the application of a reducing
agent, this was applied to the SDF-treated surface by means of
a cotton pellet until the surface turned dark brown. The excess
was removed by gentle blotting with another cotton pellet. After
treatment, the children were asked not to eat or drink for 1 hr. At
the start of the study, all children received a free tube of fluori-
dated toothpaste and a toothbrush. Children with painful or
abscessed teeth were offered free extraction.
The dentist (RY) and a dental therapist (DL) performed the
baseline examinations prior to treatment, while follow-up exam-
inations were carried out at 6, 12, and 24 mos by the dental
therapist (DL), blind to the children’s treatment group assign-
ment. All examinations were carried out in the school courtyard
or in a well-lit classroom. The children were positioned supine
on a bench or table and examined by the dental examiners using
LED headlamps and mouth mirrors. Cotton pellets were used to
dry the teeth and wipe away gross debris. A sharp sickle probe
was used, with the tip gently passed over the entire surface of
the cavity to detect and confirm visual evidence of caries. Pits
and fissures were not probed, to prevent the risk of fissure dam-
age. Ten percent of the schoolchildren were re-examined to
determine examiner reproducibility. Each surface of the tooth
was classified according to the criteria given in Table 1. For the
follow-up evaluation, surfaces with arrested caries were defined
as surfaces with active cavitated caries (d2) at baseline that
changed into surfaces with arrested cavitated caries. Surfaces
with new caries were defined as sound or ‘initial caries’ surfaces
at baseline that changed into surfaces with active or arrested
cavitated caries or restored surfaces.
Estimation of the sample size was based on the expected
number of arrested caries surfaces. Power of the study was
fixed at 80% (β = 0.20), with α = 0.05 as the significance level.
On the basis of a difference of 1 in mean number of arrested
caries surfaces between groups, and a standard deviation of 3.5
(Llodra et al., 2005), the sample size was estimated to be
around 160. When the expected dropout rate over a 2-year
period was taken into account, the sample size was increased to
240 per group.
The data were entered into a computer and analyzed with
SAS 9.1 software. Chi-square tests were applied to determine
differences in the distribution of children’s ethnic background,
parents’ education and occupation, report of oral pain, brushing
habits, and use of fluoridated toothpaste among the four groups.
Chi-square tests were also used to determine differences in num-
ber of dropouts among the four groups. Differences found in the
data of the four groups regarding mean age, dmft, ds, d2s, mean
number of non-vital teeth, and mean number of surfaces with
arrested caries were tested with analysis of variance (ANOVA).
In case of a significant ANOVA test, differences between groups
were tested with Student’s t test. Inter-examiner reproducibility
at the tooth-surface level at baseline and intra-examiner repro-
ducibility at follow-up examinations were measured by the
Since inter-examiner consistency in scoring ‘initial caries’ at
baseline was only 47%, the ‘initial caries’ scores were collapsed
Table 1. Diagnostic Criteria for the Classification of the Tooth Surface
Caries Criteria Description
Sound Surfaces not soft to the touch with a sharp sickle
probe, but they may be discolored.
Initial caries Localized enamel breakdown (microcavity) in
opaque or discolored enamel.
As d2, but the wall and floor of the cavity were
hard and could not be penetrated by the sharp
with no decay
A permanent restoration is present, and there is no
caries anywhere on the surface.
Non-vital tooth Abscessed/pulpally involved teeth with abnormal
Missing due to
Teeth extracted due to caries. Included in this
category were teeth having more than two-thirds
(2/3) of the crown destroyed by caries and the
646 Yee et al. J Dent Res 88(7) 2009
into the ‘sound’ score category prior to the analysis of the data.
After this adjustment, the unweighted inter-examiner Kappa at
baseline examination was 0.81, and intra-examiner Kappas were
0.80 (RY) and 0.81 (DL), while intra-examiner unweighted
Kappas at six-month, one-year, and two-year follow-up exami-
nations were 0.85, 0.86, and 0.93, respectively.
Although the children were randomized over the four
groups, the mean age of the children treated with 38% SDF
only was significantly lower than the mean age of children in
the other groups (Table 2). At baseline, 66% of the children
reported brushing once a day with fluoride toothpaste. The
children’s ethnic background, parents’ education and occupa-
tion, report of oral pain, brushing habits, and use of fluoridated
toothpaste were similarly distributed in the 4 groups (X2 test,
p > 0.05). The caries data and the mean number of non-vital
teeth of the four groups at baseline are shown in Table 2. Since
the numbers of missing teeth due to caries and filled teeth were
very low, the dmft was almost exclusively formed by the ‘d’
component. No statistically significant differences were found
in the caries data among the four groups at baseline (ANOVA
p > 0.05).
The numbers of children who were examined at 6, 12, and 24
mos were 908, 768, and 634, respectively, resulting in dropout
rates of 7%, 21%, and 35%. The dropout rates in the four groups
did not differ statistically significantly (X2 test, p > 0.05). The
parents’ educational level, ethnicity, brushing habits, and use of
fluoride toothpaste did not differ for the dropouts and for the
children remaining in the study at 24 mos (X2 test, p > 0.05).
There were also no differences in baseline caries parameters
between children lost to the follow-up and those remaining in
the study at 24 mos (ANOVA p > 0.05).
At 6, 12, and 24 mos, the mean number of arrested carious
surfaces was significantly higher in the two groups treated
with 38% SDF than in the 12% SDF and control groups
(Table 3). The difference observed at 6 mos decreased over 24
mos, but remained statistically significant. There was no sig-
nificant difference in the mean number of arrested carious
surfaces between the 38% SDF and the 38% SDF + tannic
acid groups, or between the 12% SDF and the control groups
throughout the 24-month study period. There was also no
significant difference between the groups in the mean number
of non-vital teeth and the mean number of exfoliated surfaces
at any time.
This is the first clinical trial to evalu-
ate the effectiveness of a one-time
application of SDF with 2 different
concentrations of SDF and the effect
of a reducing agent. The use of a
reducing agent such as 10% stan-
nous fluoride (Craig et al., 1981)
has been advocated to accelerate the
deposition of silver phosphate,
which results in the instantaneous
black discoloration of the area. This
indicates that a successful reaction
has occurred and minimizes the risk of SDF being washed away or
contaminated by saliva. However, stannous fluoride is difficult to
obtain in low-income countries, and tannic acid from boiled tea has
been suggested as an inexpensive substitute.
The large proportion of dropouts from the study (35%) at 24
mos was due to a school closing and to the mobility of the par-
ents. This did not have any effect on the results, since the 4
groups, as well as the dropouts and those remaining in the study,
were similar in all other respects.
The results of this study support the null hypothesis for the
application of 12% SDF, in that this agent had no significant
effect on arresting caries. For a single application of 38% SDF, the
hypothesis was rejected. A single application of 38% SDF, with or
without the use of tea as a reducing agent, was significantly more
effective in arresting dental caries in both the anterior and poste-
rior primary dentitions of young children than 12% SDF or no
application (control). The arresting caries effect of 38% SDF
decreases slowly over time. A single application of 38% SDF was
sufficient to prevent only 50% of the arrested surfaces at 6 mos
from reverting to active lesions again over 24 mos. The tannic
acid from boiled tea does not appear to have any significant addi-
tional effect on arresting caries compared with 38% SDF alone.
The advantages and disadvantages of this ACT approach
have been elucidated in previous studies (Chu et al., 2002;
Llodra et al., 2005). The black discoloration of the carious den-
tin after SDF treatment is probably the most notable undesirable
side-effect. This staining may be eliminated by the application of
potassium iodide (KI) after the SDF application (Knight et al.,
2005); however, the clinical effectiveness of SDF/KI remains to
be evaluated. Some concerns have also been raised over dental
fluorosis and accidental toxic overdose from the routine use of
40% silver fluoride, which has the same mode of action as SDF
(Gotjamanos and Afonso, 1997; Gotjamanos, 1997). Although
these concerns have been refuted (Neesham, 1997), the less-
frequent application of lower concentrations of SDF might help
to alleviate such concerns. In this current study, there were no
adverse effects observed or complaints from either parents or the
children concerning the SDF treatment.
The outcomes of this 24-month SDF study on both anterior
and posterior primary teeth show that: a single spot application
of 38% SDF is effective in arresting caries lesions, but the
effectiveness decreases over time; tannic acid from tea confers
no additional benefit; and 12% SDF is not effective.
Table 2. Mean Age, Mean dmft, Mean Number of Decayed, Missing and Filled Teeth, Mean Number of
Cavitated Active and Arrested Caries Surfaces, Mean Number of Cavitated Active Caries (d2) Surfaces,
and Mean Number of Non-vital Teeth of Children at Baseline According to the Different Groups
(yrs, SD) dmft (SD) ds* (SD) d2s (SD)
38% SDF 243 5.0 (1.2)a4.5 (3.1) 7.9 (7.6) 6.6 (6.4) 0.3 (1.5)
38% SDF + tea 249 5.3 (1.2)b4.7 (4.7) 8.3 (8.5) 7.2 (7.6) 0.3 (1.6)
12% SDF 243 5.3 (1.2)b4.5 (4.6) 8.0 (7.9) 6.8 (7.0) 0.3 (1.5)
Control 241 5.3 (1.2)b4.6 (4.7) 8.0 (8.5) 6.6 (7.3) 0.3 (1.6)
All groups 976 5.2 (1.2)b4.6 (4.3) 8.0 (8.1) 6.8 (7.1) 0.3 (1.6)
a<b P < 0.01.
*ds = Surfaces with cavitated active (d2) + arrested caries.
J Dent Res 88(7) 2009 Efficacy of Silver Diamine Fluoride 647
Arresting Caries Treatment (ACT) with a single spot applica-
tion of 38% SDF provides an alternative where restorative
treatment for primary teeth is not an option.
This study was financially supported by the United Mission to
Nepal Oral Health Programme and by the WHO Collaborating
Centre for Oral Health Care Planning and Future Scenarios,
Radboud University Nijmegen Medical Centre, College of
Baelum V, van Palenstein Helderman W, Hugoson A, Yee R, Fejerskov O
(2007). A global perspective on changes in the burden of caries and
periodontitis: implications for dentistry. J Oral Rehabil 34:872-906.
Bedi R, Sardo-Infirri J (1999). The root cause–a proposal for action.
London, UK: FDI World Dental Press Ltd.
Chu CH, Lo EC, Lin HC (2002). Effectiveness of silver diamine fluoride
and sodium fluoride varnish in arresting dentin caries in Chinese pre-
school children. J Dent Res 81:767-770.
Craig GG, Powell KR, Cooper MH (1981). Caries progression in primary
molars: 24-month results from a minimal treatment programme.
Community Dent Oral Epidemiol 9:260-265.
Edelstein BL (2006). The dental caries pandemic and disparities problem.
BMC Oral Health 6(Suppl 1):2.
Gotjamanos T (1997). Safety issues related to the use of silver fluoride in
paediatric dentistry. Aust Dent J 42:166-168.
Gotjamanos T, Afonso F (1997). Unacceptably high levels of fluoride in
commercial preparations of silver fluoride. Aust Dent J 42:52-53.
Knight GM, McIntyre JM, Craig GG, Mulyani, Zilm PS, Gully NJ (2005).
An in vitro model to measure the effect of a silver fluoride and potas-
sium iodide treatment on the permeability of demineralized dentine to
Streptococcus mutans. Aust Dent J 50:242-245.
Llodra JC, Rodríguez A, Ferrer B, Menardia V, Ramos T, Morato M (2005).
Efficacy of silver diamine fluoride for caries reduction in primary teeth
and first permanent molars of schoolchildren: 36-month clinical trial. J
Dent Res 84:721-724.
Lo EC, Chu CH, Lin HC (2001). A community-based caries control program
for pre-school children using topical fluorides: 18-month results. J Dent
Neesham DC (1997). Fluoride concentration in AgF and dental fluorosis
(letter to editor). Aust Dent J 42:268-269.
Nishino M, Yoshida S (1969). Clinical effects of diamine silver fluoride on
caries and on pulp of deciduous teeth [in Japanese]. Jpn J Pedodont
Nishino M, Yoshida S, Sobue S, Kato J, Nishida M (1969). Effect of topi-
cally applied ammoniacal silver fluoride on dental caries in children. J
Osaka Univ Dent Sch 9:149-155.
Yamaga R, Yokomizo 1 (1969). Arrestment of caries of deciduous teeth with
diamine silver fluoride. Dental Outlook 33:1007-1013.
Table 3. Mean Number of Active Cavitated Surfaces at Baseline that Changed at Follow-up Examination into Surfaces with Arrested Cavitated Caries
and Into Mean Number of Surfaces Lost Due to Exfoliation and Mean Number of Non-vital Teeth at 6-, 12-, and 24-month Follow-up Examinations
6 Mos 12 Mos 24 Mos
38% SDF 223 4.2(0.3)a0.4(0.1) 0.0(0.0) 189 3.4(0.3)a0.6(0.1) 0.0(0.0) 157 2.1(0.3)a0.4(0.1) 0.4(0.1)
38% SDF + tea 233 4.5(0.4)a0.4(0.1) 0.0(0.0) 189 4.1(0.4)a0.6(0.1) 0.0(0.0) 156 2.2(0.3)a0.4(0.1) 0.4(0.1)
12% SDF 228 2.3(0.2)b0.4(0.1) 0.0(0.0) 188 1.7(0.3)b 0.6(0.1) 0.1(0.0) 156 1.5(0.3) 0.3(0.1) 0.4(0.1)
Control 223 1.6(0.2)b 0.5(0.1) 0.0(0.0) 187 1.3(0.2)b0.7(0.1) 0.1(0.0) 155 1.0(0.2)b0.5(0.1) 0.4(0.1)
a>b P < 0.001 a>b P < 0.001 a>b P < 0.01